Serious infections during anti-TNFα treatment in rheumatoid arthritis patients

Autoimmunity Reviews 8 (2009) 266–273

Contents lists available at ScienceDirect

Autoimmunity Reviews j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / a u t r ev

Serious infections during anti-TNFα treatment in rheumatoid arthritis patients☆ Ennio Giulio Favalli a,⁎, Francesca Desiati a, Fabiola Atzeni b, Piercarlo Sarzi-Puttini b, Roberto Caporali c, Francesca Bobbio Pallavicini c, Roberto Gorla d, Matteo Filippini d, Antonio Marchesoni a a b c d

UOC Day Hospital of Rheumatology, G. Pini Orthopedic Institute, University of Milan, Milan, Italy Rheumatology Unit, L. Sacco University Hospital, Milan, Italy IRCCS Policlinico S. Matteo, Pavia, Italy Rheumatology and Immunology Unit, Spedali Civili di Brescia, Brescia, Italy

a r t i c l e

i n f o

Article history: Received 12 October 2008 Accepted 1 November 2008 Available online 18 November 2008 Keywords: Infliximab Etanercept Adalimumab Rheumatoid arthritis Anti-TNFα Infection Safety

a b s t r a c t The objective was to estimate the incidence of serious infections in the patients treated with anti-TNFα agents for rheumatoid arthritis (RA) recorded in the Lombardy Rheumatology Network (LORHEN) registry. The study inclusion criteria were met by 1064 of the 1114 patients with long-standing RA, 519 treated with infliximab, 303 with adalimumab, and 242 with etanercept; their mean age was 55.8 years and the mean duration of RA 9.4 years. Seventy-three patients (6.9%) experienced a total of 74 serious infections, an incidence rate for all treatment courses of 35.9 per 1000 patient-years (95% confidence interval [95% CI] 27.66– 44.13). Most were lower respiratory tract (34.2%) or skin and soft tissue infections (20.5%). Of the 1064 patients, the 790 treated with anti-TNFα after March 2002 underwent screening tests for LTBI; five patients developed active tuberculosis. Three patients died of septic shock. The type of anti-TNFα agent did not seem to affect the incidence or site of the infections. Both univariate and multivariate analyses identified age at the start of anti-TNFα treatment (p = 0.008), baseline erythrocyte sedimentation rate ([ESR] p = 0.014), and the concomitant use of corticosteroids (p = 0.029) as significant predictors of infections. There was no statistically significant difference in risk between the anti-TNFα agents. © 2008 Elsevier B.V. All rights reserved.

Contents 1. 2.

3.

Introduction . . . . . . . . . . . . . . . . . Patients and methods . . . . . . . . . . . . . 2.1. The LORHEN registry . . . . . . . . . . 2.2. Patient characteristics . . . . . . . . . 2.3. Identification and classification of serious 2.4. Statistical analyses . . . . . . . . . . . Results . . . . . . . . . . . . . . . . . . . . 3.1. Study population. . . . . . . . . . . . 3.2. Incidence rates of serious infections. . . 3.3. Incidence rate and risk of tuberculosis .

. . . . . . . . . . . . . . . . . . . . infections . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

☆ All of the authors have received consultancy fees or Congress invitations from Schering-Plough, Wyeth, and Abbott. ⁎ Corresponding author. UOC Day Hospital of Rheumatology, G. Pini Orthopedic Institute, Via G. Pini 9, 20122 Milano, Italy. E-mail address: [email protected] (E.G. Favalli). 1568-9972/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.autrev.2008.11.002

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

267 267 267 267 268 268 268 268 269 270

E.G. Favalli et al. / Autoimmunity Reviews 8 (2009) 266–273

3.4. Risk factors associated with 4. Discussion. . . . . . . . . . . . 5. Conclusions . . . . . . . . . . . Take-home messages . . . . . . . . . References . . . . . . . . . . . . . .

infection. . . . . . . . . . . . . . . . . . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

1. Introduction More than 1,500,000 patients throughout the world have received anti-TNFα agents for all therapeutic indications since their introduction in 1999, including many who have experienced a dramatic improvement in severe rheumatoid arthritis (RA). Randomised clinical trials (RCTs) of the three currently licensed agents (the monoclonal antibodies infliximab and adalimumab, and the soluble TNFα receptor etanercept) have shown that they are effective in reducing inflammatory activity and limiting joint destruction in patients with active early and long-standing RA [1–6]. However, they have also raised a number of safety concerns of which, given the major role of TNFα in host defence mechanisms, one of the most important is an increased risk of infection. Infectious adverse events are common in patients with rheumatic diseases [7] and it has been reported that the rate of infections in the RA population is nearly twice as high as that observed in matched non-RA controls [8], which may be related to the underlying disease itself (alterations in immunological functions, disability) or the drugs used to treat it (especially immunosuppressants drugs and corticosteroids). The question that needs to be answered is whether anti-TNFα treatment further increases that risk. The greatest concern is tuberculosis because the use of TNFα antagonists is accompanied by an increased susceptibility to active tuberculosis (ATB) or the reactivation of latent tuberculosis infection (LTBI) [9], which should be considered characteristic of the drug class [10]. However, health authorities published recommendations for screening patients with LTBI who were candidates for TNFα antagonist treatment in 2002, and the subsequent decrease in the number of reports of new ATB cases suggests that the problem of LTBI is now under control [11]. Nevertheless, there are still concerns relating to other infective conditions as well as a lack of information concerning the long-term safety of the daily use of biological agents. There have been a number of post-marketing surveillance and other reports indicating the occurrence of serious infections during the use of TNFα blockers (including opportunistic infections such as candidosis, listeriosis, nocardiosis, histoplasmosis, aspergillosis and pneumocystosis) [12–15], but the results of RCTs do not consistently demonstrate any significant increase in the incidence of opportunistic or serious infections in patients treated with anti-TNFα agents in comparison with those observed in patients receiving methotrexate plus placebo [1–6]. However, these RCTs may have had insufficient power to detect such an increase because their designs and stringent inclusion/exclusion criteria may limit enrolment to patients at low risk of infection. Moreover, questions concerning the long-term safety of therapy with biological agents cannot be addressed by short-term RCTs not only because of

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

267

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

271 271 272 272 272

their short duration, but also because they recruit too few patients to detect rare events. For these reasons, large population-based registries that allow long-term follow-up are increasingly being used to investigate the incidence of adverse events in RA patients exposed to anti-TNFα drugs [16–20]. However, as no national pharmacovigilance registry has yet been planned in Italy, we conducted a prospective population-based cohort study of RA patients enrolled in four major Lombardy Rheumatology Units (the LOHREN Registry) [21]. 2. Patients and methods 2.1. The LORHEN registry LORHEN is a regional population-based registry that was established with the aim of examining the efficacy and safety of TNFα blockers in the everyday treatment of RA. LORHEN contains a cohort of all of the RA patients treated with antiTNFα agents in four major Rheumatology Units in Lombardy (Italy) since 1999, when they first became available for compassionate use [21]. This analysis covers the first 36 months from the date of the first administered dose (median follow-up: 23.22 months for etanercept, 21.26 months for adalimumab, 29.88 months for infliximab, and 24.21 months for the patients as a whole). Total exposure to anti-TNF treatment was 2033.62 person/ years (430.09 for etanercept, 524.00 for adalimumab, and 1079.53 for infliximab). 2.2. Patient characteristics Patients were included in the study cohort if they had a diagnosis of RA (ACR criteria), were ≥18 years old, and had received at least one infusion or filled a prescription for one TNFα blocker (e.g. etanercept, infliximab, or adalimumab). All of the patients were treated in accordance with the Italian Society of Rheumatology guidelines for the use of anti-TNFα agents: a diagnosis of RA (ACR criteria), a failure to respond to at least one course of combination therapy with full-dose traditional disease modifying anti-rheumatic drugs (DMARDs), one of which should always be methotrexate (MTX) unless contraindicated, and active disease as defined by a Disease Activity Score on 28 joints (DAS28) of N3.5. The exclusion criteria were active infection, a history of malignancy or a pre-malignant condition, class III/IV congestive heart failure and demyelinating disorders. During the observational period, each patient could have been treated with more than one TNFα antagonist and therefore could have received various treatment courses. The recorded baseline data were demographic characteristics (including gender and date of birth), disease severity and duration, drug therapy (including the number of previous

268

E.G. Favalli et al. / Autoimmunity Reviews 8 (2009) 266–273

DMARDs) and comorbidities, and the following data were systematically collected throughout the treatment period: 1) All anti-TNFα treatments, including the type of agent, starting date, doses, treatment duration (and, if applicable, the date of discontinuation together with the reason), and concomitant treatments and doses (DMARDs and/or corticosteroids); 2) Adverse events, date(s) of occurrence, and comorbidities. There was initially no specific indication concerning the risk of tubercular infections but, when the role of anti-TNFα agents in reactivating LTBI became clear, the Units established guidelines regarding the prevention of LTBI in candidates for anti-TNFα treatment. The recommendations [22] included nine months of isoniazid (INH) treatment for all patients with: 1) a tuberculin skin test (TST) result of ≥5 mm; 2) a chest radiograph showing evidence of previous tuberculosis; and 3) previous exposure to or partially treated active tuberculosis. TST results, chest radiology findings and LTBI prophylactic therapy were recorded from March 2002. 2.3. Identification and classification of serious infections All physicians documented episodes of infection requiring medical care that occurred during biological therapy; only the infectious events recorded in medical charts or during outpatient visits were available. The minimum at-risk window to maintain a biologically plausible link between medication exposure and an infectious adverse event should extend from the beginning of anti-TNFα treatment to the date of its discontinuation. However, given the administration schedule of the agents (especially the fact that infliximab infusions are given several weeks apart), and their pharmacokinetic and pharmacodynamic characteristics, we decided to define the “anti-TNFα stop date” as the date of the first missed scheduled dose [23]. All of the recorded infections were categorised as “serious” (defined as life-threatening, requiring hospitalisation and/or intravenous antibiotic therapy, or leading to significant disability/incapacity or a comparable significant risk) or “non-serious”, defined as any other infectious episode. These definitions were similar to those used in the phase III anti-TNF clinical trials. Given the observational, everyday practice design of the LORHEN registry, the data concerning non-serious infections were often missing or incomplete and were therefore not included in the final analysis. Nine infectious patterns were identified on the basis of the site of infection: bacteremia/septicemia; lower respiratory tract infections, including pneumonia and bronchitis; skin and soft tissue infections, including cellulitis, abscesses, wound infections and herpes zoster; intra-abdominal infections (which could be included on the basis of physician diagnosis and relevant clinical findings), including acute cholecystitis, ascending cholangitis, suppurative appendicitis, peritonitis and gastroenteritis; urinary tract infections, including pyelonephritis and urosepsis (the isolation of N100.000 colony-forming units/ ml of urine in the presence of clinical features); joint and bone infections (included osteomyelitis and septic arthritis); neurological infections; and “other infections”, including eye and hear infections, sinusitis, genital infections, acute hepatitis and

pericarditis. The infections were also categorised on the basis of the causative micro-organism as bacterial, viral, mycobacterial, fungal, parasitic, or “not defined”. The infectious patterns were descriptively analysed for all patients during all treatment courses and by anti-TNFα antagonist. 2.4. Statistical analyses The primary endpoint was the occurrence of a first serious infection in the cohort of RA patients starting any anti-TNFα therapy, and so the incidence rates and predictors were analysed using an intention-to-treat (ITT) approach. The patients were considered at risk of serious infection from the date of the start of the first anti-TNFα drug (baseline) to the date of the first occurrence of the event or the last date of follow-up within 36 months, regardless of any treatment discontinuation or switch. The incidence rates (IRs) and Poisson's 95% confidence intervals (CIs) were calculated from the observed number of events and the number of person-years at risk; the IRs and 95% CIs of all infections and the TB cases were estimated in the whole population eligible for this analysis and on the basis of specific baseline characteristics. The events occurring during the first treatment course (OT) were also described. Multivariate analyses based on Cox regression models were used to identify the independent predictors of the primary endpoint. The relative risks of developing a serious infection are given as hazard ratios and 95% CIs for all of baseline variables included in the models. The Kruskal–Wallis, Mann–Whitney non-parametric test, Pearson's chi-squared test and Fisher's exact test were used to evaluate the differences in continuous and categorical variables between the three treatment groups (infliximab, adalimumab, etanercept) and between patients with positive and negative TST results. All of the analyses were performed using the statistical software package SPSS 15 for Windows, with the significance level set at 0.05. 3. Results 3.1. Study population Since October 1999, a total of 1114 RA patients have received at least one anti-TNFα dose, of whom 833 (79.9%) have received only a first-line (22.4% etanercept, 29.8% adalimumab, and 47.8% infliximab), 237 (17%) a second-line (65.8% etanercept, 28.7% adalimumab, and 5.5% infliximab), and 44 a third-line (22.7% etanercept, 68.2% adalimumab, and 9.1% infliximab), for a total of 1395 treatment courses. Fifty patients were lost to follow-up during the first six months of treatment, and so only 1064 patients (83.2% women) were considered for the safety analysis. Table 1 summarises the characteristics of the study cohort. At the time of starting their first anti-TNFα drug, their mean age was 55.84 years and the mean duration of RA was 9.44 years; 76.1% were positive for rheumatoid factor (RF). Over 60% of the patients had comorbidities, including diabetes mellitus (1.7%), chronic obstructive pulmonary disease or asthma (2.2%),

E.G. Favalli et al. / Autoimmunity Reviews 8 (2009) 266–273

269

Table 1 Baseline demographic and clinical data

Age (years) Females Males Disease duration (years) Disease duration b5 years DAS28 DI-HAQ RF positive Comorbidity Chronic lung disease Diabetes mellitus Leukopenia Chronic hepatitis Hypertension Dyslipidemia Other Functional class III/IV Methotrexate therapy Corticosteroid therapy

All patients

Infliximab

Adalimumab

Etanercept

1064

519 (48.8%)

303 (28.5%)

242 (22.7%)

55.84 ± 12.96 885 (83.2%) 179 (16.8%) 9.44 ± 7.29 324 (30.5%) 5.90 ± 0.97 1.46 ± 0.61 810 (76.1%) 657 (61.7%) 61 (5.7%) 18 (1.7%) 2 (0.2%) 39 (3.7%) 274 (25.8%) 59 (5.5%) 287 (27.0%) 317 (29.8%) 899 (84.5%) 896 (84.2%)

55.72 ± 12.07 423 (81.5%) 96 (18.5%) 9.28 ± 7.00 160 (30.8%) 6.01 ± 0,94 1.61 ± 0.61 417 (80.3%) 316 (60.9%) 24 (4.6%) 12 (2.3%)

56.07 ± 13.11 258 (85.1%) 45 (14.9%) 9.56 ± 7.90 106 (35.0%) 5.68 ± 0.96 1.25 ± 0.54 214 (70.6%) 179 (59.1%) 18 (5.9%) 2 (0.7%) 2 (0.7%) 8 (2.6%) 82 (27.1%) 18 (5.9%) 77 (25.4%) 63 (20.8%)/226 (74.6%) 226 (74.6%) 233 (76.9%)

55.81 ± 14.57 204 (84.3%) 38 (15.7%) 9.63 ± 7.11 58 (24.0%) 5.93 ± 1.02 1.41 ± 0.63 179 (74.0%) 162 (66.9%) 19 (7.9%) 4 (1.7%)

11 (2.1%) 124 (23.9%) 31 (6.0%) 151 (29.1%) 186 (35.8%) 499 (96.1%) 459 (88.4%)

p

20 (8.3%) 68 (28.1%) 10 (4.1%) 59 (24.4%) 68 (28.1%) 174 (71.9%) 204 (84.3%)

ns ns ns ns 0.020 0.000 0.000 0.001 ns ns ns – 0.000 ns ns ns 0.000 0.000 0.000

Continuous variables are expressed as mean values ± SD. ns: not statistically significant.

pulmonary interstitial fibrosis (3.8%), leukopenia (0.2%), and chronic hepatitis (3.7%). The median number of previous DMARDs was three per patient, the vast majority of whom had previously received MTX. In combination with the biological drug, 84.2% of the patients received steroids (mean dose 4.92 mg/day) and 88.0% at least one DMARD (84.5% methotrexate at a median dose of 12.35 mg/week). 3.2. Incidence rates of serious infections Seventy-three patients (6.9%) experienced a total of 74 serious infectious events, which were life-threatening in 12, led to death in four, required hospitalisation and/or intravenous antibiotic therapy in 27, or led to significant disability/

incapacity or a comparable significant risk in 31. The IR for all treatment courses was 35.9 (95% CI: 27.66–44.13) per 1000 patient-years, and seemed to increase with age. Table 2 shows the overall rates, infectious patterns and responsible micro-organisms. Most of the infections (34.3% of the total) involved the lower respiratory tract (IR 12.29 [95% CI 7.47–17.11] per 1000 patient-years), including 14 cases of pneumonia (19.2%, including one case of Legionella pneumophila pneumonia), 11 cases of bronchitis (15.1%), and one case of pulmonary tuberculosis. Skin and soft tissue infections were also frequent (15 cases, IR 7.38 [95% CI 3.64–11.11] per 1000 patient/years), and included eight Herpes zoster infections (11.0%), five bacterial infections (6.9%), and one subcutaneous disseminated tubercular infection. Finally,

Table 2 Frequency of serious infections by infection site and responsible micro-organism

Any serious infection Site of infections Skin and soft tissue Lower respiratory tract Urinary tract Osteoarticular Neurological Intra-abdominal Sepsis Other Micro-organisms Bacterial Viral Mycobacterial Fungal Parasitic Not defined a b

All patients

Infliximab

Adalimumab

Etanercept

1064

519 (48.8%)

303 (28.5%)

242 (22.7%)

n (%) a

IR (95% CI) b

n (%) a

IR (95% CI) b

n (%) a

IR (95% CI) b

n (%) a

IR (95% CI) b

73 (6.9)

35.90 (27.66–44.13)

42 (8.1)

38.91 (27.14–50.67)

20 (6.6)

38.17 (21.44–54.90)

11 (4.5)

25.58 (10.46–40.69)

15 (1.4) 25 (2.3) 6 (0.6) 7 (0.7) 1 (0.1) 7 (0.7) 3 (0.3) 9 (0.8)

7.38 (3.64–11.11) 12.29 (7.47–17.11) 2.95 (0.59–5.31) 3.44 (089–5.99) 0.49 (0.00–1.46) 3.44 (0.89–5.99) 1.48 (0.00–3.14) 4.43 (1.53–7.32)

11 (2.1) 16 (3.1) 4 (0.8) 4 (0.8) 0 2 (0.4) 1 (0.2) 4 (0.8)

10.19 (4.17–16.21) 14.82 (7.56–22.08) 3.71 (0.07–7.34) 3.71 (0.07–7.34) 0 1.85 (0.00–4.42) 0.93 (0.00–2.74) 3.71 (0.07–7.34)

3 (1.0) 7 (2.3) 0 0 0 3 (1.0) 2 (0.7) 5 (1.7)

5.73 (0.00–12.20) 13.36 (3.46–23.25) 0 0 0 5.73 (0.00–12.20) 3.82 (0.00–9.11) 9.54 (1.18–17.91)

1 (0.4) 2 (0.8) 2 (0.8) 3 (1.2) 1 (0.4) 2 (0.8) 0 0

2.33 (0.00–6.88) 4.65 (0.00–11.09) 4.65 (0.00–11.09) 6.98 (0.00–14.87) 2.33 (0.00–6.88) 4.65 (0.00–11.09) 0 0

41 (3.9) 12 (1.1) 5 (0.5) 2 (0.2) 0 13 (1.2)

20.16 (13.99–26.33) 5.90 (2.56–9.24) 2.46 (0.30–4.61) 0.98 (0.00–2.35) 0 6.39 (2.92–9.87)

23 (4.4) 9 (1.7) 3 (0.6) 1 (0.2) 0 5 (1.0)

21.31 (12.60–30.01) 8.34 (2.89–13.78) 3.71 (0.07–7.34) 0.93 (0.00–2.74) 0 4.63 (0.57–8.69)

9 (3.0) 3 (1.0) 1 (0.3) 0 0 7 (2.3)

17.18 (5.95–28.40) 5.73 (0.00–12.20) 1.91 (0.00–5.65) 0 0 13.36 (3.46–23.25)

9 (3.7) 0 1 (0.4) 1 (0.4) 0 1 (0.4)

20.93 (7.25–34.60) 0 2.33 (0.00–6.88) 2.33 (0.00–6.88) 0 2.33 (0.00–6.88)

Percentage of all infections. Incidence rate (IR): number of events per 1000 patient-years (95% confidence interval).

270

E.G. Favalli et al. / Autoimmunity Reviews 8 (2009) 266–273

Table 3 TB screening tests and frequency, and characteristics of TB infections

Screening tests Performed Positive TB cases Pre-OR c Post-OR c Pulmonary Extra-pulmonary Mean period after first drug intake (days) a b c

All patients

Infliximab

Adalimumab

Etanercept

1064

519 (48.8%)

303 (28.5%)

242 (22.7%)

n (%) a

IR (95% CI) b

n (%) a

IR (95% CI) b

n (%) a

IR (95% CI) b

n (%) a

IR (95% CI) b

790 (74.2) 115 (14.6) 5 2 3 1 4 96.4

– – 246 (30–461) 323 (0–770) 212 (0–452) 49 (0–146) 197 (0–389) –

282 (54.3) 27 (9,6) 3 2 1 0 3 70

– – 278 (0–592) 375 (0–895) 183 (0–542) 0 278 (0–592) –

303 (100) 45 (14.9) 1 0 1 0 1 183

– – 191 (0–565)

205 (84.7) 43 (21.0) 1 0 1 1 0 90

– – 233 (0–688)

191 (0–565) 0 191 (0–565) –

291 (0–862) 233 (0–688) 0 –

Percentage of all infections. Incidence rate (IR): number of events per 100.000 patient-years (95% confidence interval). Cases of active TB before (pre-OR) and after (post-OR) the publication of official recommendations for managing latent TB infection.

there were three cases of septic arthritis, one of acute osteomyelitis, and three of periprosthetic septic arthritis. Two patients (one treated with infliximab and one with adalimumab) died of septic shock following severe bacterial infections subsequent to surgical procedures. Bacteria were responsible for 56.2% of the infectious events, viruses for 16.4%, fungi for 2.7%, and mycobacteria for 6.9%; no parasitic infections were recorded, but the microorganism was not identified in 17.8% of cases. Nine infections occurred during treatment with etanercept (3.71% of etanercept treatment courses), 31 during treatment with infliximab (5.97%), and 17 during treatment with adalimumab (5.61%) after a mean (±SD) duration of treatment of 15.9 ± 10.2 months. The type of anti-TNFα agent did not seem to affect the IRs or sites of infection, or the nature of the responsible micro-organism. 3.3. Incidence rate and risk of tuberculosis All of the 790 patients who started treatment with an antiTNFα agent from March 2002 underwent LTBI screening tests. Because of the different times the drugs were available on the market, all of the patients treated with adalimumab underwent a TST, 84.7% of those treated with etanercept, and only 54.3% of those treated with infliximab (Table 3). The TST results were positive in 115 patients, all of whom were treated with INH. In comparison with the patients with negative screening test results, these patients were older (58.8 ± 11.9 vs 55.7 ± 13.4 years; p = 0.02) and more frequently treated with etanercept than infliximab (37.4% vs 23.5%; p = 0.002); moreover, at the time of screening, they were less frequently receiving concomitant DMARD therapy (13.3 vs 24.3%; p = 0.004) or corticosteroids (16.3% vs 24.3%; p = 0.04). Five women had ATB (mean age 66.4 ± 6 years), corresponding to 246 cases per 100,000 patient-years (95% CI: 30– 461), none of whom received prophylactic INH therapy. Two (all treated with infliximab) did not undergo LTBI screening, and three (respectively treated with infliximab, etanercept, and adalimumab) were negative upon testing. One patient had active pulmonary infection, three extra-pulmonary infections (tubercular cystitis, tonsillitis, and arthritis), and one disseminated infection. All were treated with antituberculosis drugs, four were hospitalised, and none died.

Three of these patients developed ATB respectively 43, 68 and 98 days (mean 70 days) after starting treatment with infliximab; one after 183 days of treatment with adalimumab; and the last, initially treated with infliximab for two years (discontinued because of lack of efficacy) was diagnosed as having ATB during the second anti-TNFα treatment course with etanercept (90 days after the first intake of the drug).

Table 4 Univariable and multivariable predictors of serious infections p Demographics Age (per 1 year) 0.002 Males 0.096 RA characteristics Disease duration 0.194 N5 years Functional 0.065 class III/IV No. of previous 0.050 DMARDs N 2 HAQ 0.076 RF positive 0.376 ESR 0.012 DAS28 0.103 Anti-TNF treatment Etanercept Adalimumab 0.285 Infliximab 0.237 Concomitant treatment Methotrexate 0.865 Prednisone 0–5 mg 0.297 Prednisone N 5 mg 0.025 Comorbidity factors 0.055 Lung chronic 0.506 diseases COPD/Asthma 0.241 Pulmonary fibrosis 0.303 Diabetes 0.934 Dislipidemy 0.920 Chronic hepatitis 0.393 Hypertension 0.003 Leukopenia Other 0.792 a

HRa

95% CIb

p

AHRc 95% CIb

1.03 1.01 1.05 0.52 0.24 1.13

0.008 1.03 0.362 0.68

1.01 1.06 0.30 1.55

1.44

0.83 2.51

0.670 1.14

0.62 2.08

1.55

0.97 2.47

0.387 1.28

0.73 2.25

1.68

1.00

2.82 0.118

1.58

0.89 2.79

1.40 1.34 1.01 1.23

0.97 0.70 1.00 0.96

2.02 2.55 1.02 1.58

0.65 1.08 1.02 0.75

0.37 0.55 1.00 0.40

1 1.49 1.49

1 0.72 3.12 0.182 1.73 0.77 2.90 0.304 1.48

0.77 3.87 0.70 3.14

1.07 1.53 2.69 1.66 1.33

0.49 0.69 1.13 0.99 0.58

2.34 3.38 6.41 2.77 3.06

0.730 0.393 0.029 0.379 0.591

1.17 1.43 2.89 1.30 1.27

0.49 0.63 1.12 0.72 0.53

2.80 3.28 7.48 2.34 3.08

2.00 1.61 0.92 1.05 0.42 2.05 – 0.93

0.63 0.65 0.13 0.38 0.06 1.29

6.34 4.00 6.62 2.89 3.05 3.27

0.301 0.508 0.621 0.564 0.329 0.071

1.94 1.39 0.60 0.73 0.37 1.74 – 0.856 1.07

0.55 0.52 0.08 0.25 0.05 0.95

6.77 3.71 4.64 2.13 2.75 3.19

HR: hazard ratio. b 95% CI: 95% confidence interval. c AHR: adjusted hazard ratio.

0.55 1.57

0.145 0.831 0.014 0.372

1.16 2.12 1.03 1.42

0.53 2.13

E.G. Favalli et al. / Autoimmunity Reviews 8 (2009) 266–273

3.4. Risk factors associated with infection Univariate analysis (Table 4) revealed only three statistically significant predictors of an increased risk of infection: age at the time of starting biological drug treatment (p = 0.002), the baseline erythrocyte sedimentation rate ([ESR] p = 0.012), and the concomitant use of corticosteroids (p = 0.025). Gender, disease duration, concomitant DMARD therapy, positive rheumatoid factor, and poor functional capacity did not seem to be predictive, and none of the three TNFα blockers seemed to be associated with an increased risk. Multivariate models confirmed that the use of corticosteroids at a daily dose of N5 mg (adjusted hazard ratio [AHR] 2.89 [95% CI: 1.12–7.48]) was a strong and statistically significant predictor of infection (p = 0.029). The other factors independently associated with an increased risk of serious infections were increasing age (AHR 1.03/year [95% CI 1.01–1.06]) and ESR (AHR 1.02 per every mm/h increase [95% CI: 1.00–1.03]), but not the use of infliximab (AHR 1.48 [95% CI: 0.70–3.14]) or adalimumab (AHR 1.73 [95% CI: 0.77–3.87]) in comparison with etanercept (p = 0.304 and p = 0.182, respectively). The presence of comorbidity did not predict infectious adverse events among the patients who were diabetic (AHR 0.60 [95% CI: 0.08–4.64]), those with previous pulmonary disease (AHR 1.27 [95% CI: 0.53–3.08]), or those with chronic hepatitis (AHR 0.37 [95% CI 0.05–2.75]). 4. Discussion The aim of this study, conducted under conditions of everyday practice, was to record the incidence of serious infections in RA patients treated with anti-TNFα agents, as well as all information concerning infection patterns, characteristics and risk factors, concentrating on the potential role of the individual agents. The overall rate of serious infections in our cohort (35.9 [95% CI: 27.66–44.13] per 1000 patient-years) was significantly lower than those reported in other post-marketing observational studies. Kroesen et al. [24] prospectively evaluated serious infections (defined as requiring intravenous antibiotic therapy and/or hospitalisation) in a small group of 60 RA patients treated with etanercept or infliximab, and found an IR of 181 per 1000 patient-years; similarly, in a cohort of 709 RA patients, Salliot et al. [16] compared the incidence of infections during the first TNFα blocker course with that observed in the preceding period, and found a rate of serious infections of 105 (±86.9) per 1000 patient-years; and Dixon et al. [18] have recently published data from the British Society for Rheumatology Biologics Register (7664 RA patients) showing an overall incidence of serious infections during anti-TNF treatment of 53.2 (95% CI: 48.9–57.8) per 1000 patient-years. The differences are probably due to the fact that these were prospective studies whereas ours was retrospective, which may underestimate some infections that may not have been fully reported to the treating rheumatologist. However, the frequency of serious infections in our study was similar to that reported by Maini et al. [2] and Lipsky et al. [5] in the ATTRACT study (6% for the first year of infliximab therapy, compared with 5.9% in our study), Klareskog et al. [4] in the TEMPO trial (4% for the first year of etanercept, compared with 3.7% in our study), and by Weinblatt et al. [3]

271

in the ARMADA trial (2.03 per 100 patient-years for adalimumab, compared with 3.81 per 100 patient-years in our study). As indicated in previous studies [16–19], we found that the most frequent site of serious infection was the lower respiratory tract (12.29/1000 patient-years). In line with the major theoretical role of TNFα in infectious lung diseases revealed by the high bronchoalveolar fluid TNFα levels in the infected pulmonary lobes of patients with communityacquired pneumonia [25], pneumonia was the most frequent infection in our cohort (fourteen cases, 17.8% of all reported infections). However, we found no specific correlation between the use of anti-TNFα and the risk of pneumonia. Our results are consistent with those recently reported by Wolfe et al. [26], who found a high rate of hospitalisation due to pneumonia in RA patients (14.7 [95% CI 13.1–16.4] per 1000 patient-years) regardless of anti-TNF treatment. The literature includes reports of bacterial skin infections, such as erysipelas and cellulitis [16,17], and we recorded five such cases. The key role of TNFα in cutaneous immunity has been widely demonstrated [27] and thus TNFα inhibition in the skin may decrease its sensitivity and response to infection. On the contrary, the evidence of viral infections is less convincing [28], although we recorded eight cases of Herpes zoster infection (11% of all serious infections). The infection that causes the greatest concern is tuberculosis [9,10,29,30]. In 2005, the annual IR of ATB in the Italian population was 7 per 100,000 persons [31], whereas we had five cases in our cohort, for an incidence of 246 (95% CI: 30–461) per 100,000 patient-years. However, as two of these cases occurred during the pre-screening era, the incidence rate in the group of screened patients (n = 790) was 212 (95% CI: 0–452) per 100,000 patient-years. As expected, concomitant treatment with immunosuppressive agents such as MTX and corticosteroids may influence TSTs, which were more frequently positive in the patients not receiving DMARDs or steroids. Although the risk of ATB during anti-TNF therapy seems to apply to all three available agents, it has been suggested that monoclonal antibodies carry a higher risk than etanercept because of their different effects on tubercular granuloma structure and macrophage activation [32]. The early clinical reports of tuberculosis seemed to indicate a substantially greater risk with infliximab than with etanercept [28,33], and this has recently been confirmed by the BIOBADASER registry results (infliximab IR 383 [95% CI: 159–921], etanercept 114 [28–459], and adalimumab 176 [24–1254] per 100,000 patient-years) [10]. We also recorded more cases of ATB with infliximab (n = 3) than with etanercept (n = 1) or adalimumab (n = 1), but it is important to stress that the calendar year in which the use of the drug was started may have influenced the rate of ATB. There was a national shortage of etanercept in Italy up to early 2004, and so the first TNFα blocker for most patients was infliximab, which was frequently prescribed before the need for LTBI screening became apparent, whereas adalimumab was marketed after the risk of TB had become clear and LTBI screening had been introduced. This may have led to a higher rate of ATB in adalimumab- and etanercept-treated patients, and a lower rate in patients receiving infliximab. In this regard, it is worth pointing out that two of our three cases of ATB during infliximab therapy occurred in unscreened patients, whereas the incidence of

272

E.G. Favalli et al. / Autoimmunity Reviews 8 (2009) 266–273

ATB with the three TNF antagonists was exactly the same after screening had been started. Our data confirm previous reports of an unusual pattern of TB infection in patients treated with anti-TNFα agents [18,34]: about 85% of the cases of TB in the general population are pulmonary, whereas we had four (80%) extra-pulmonary infections. Our results also confirm that TB is an early adverse event during anti-TNF therapy: the mean time after the start of treatment was 96.4 days, whereas TB occurred in only one case (treated with adalimumab) after as long as 183 days. In line with other reports [16,20], we found that the risk factors for infections were age at the time of starting the biological drug, baseline ESR, and the concomitant use of highdose corticosteroids. The role of anti-TNF agents in increasing this risk is still unclear as our findings are not consistent with those reported by other authors [16,19], who have found an increased risk of infections in patients exposed to anti-TNF drugs. Our results must be interpreted in the context of our study design, which is characterised by some strengths and some potential limitations. First of all, it is important to underline that this was a retrospective observational cohort study and not a randomised comparative trial. We collected data relating to patients receiving anti-TNF therapy, but the absence of a control group prevents us from comparing our findings with data coming from a matched cohort of patients treated with DMARDs. Moreover, we had to limit our analysis to serious infections because the data concerning non-serious infections were underreported, missing or incomplete. However, randomised clinical trials cannot adequately study infectious adverse events because of the limitations imposed by inclusion/exclusion criteria[35,36], whereas the results of our study conducted under conditions of everyday clinical practice may be more reliable. 5. Conclusions In a setting of clinical practice, we found that the overall rate of serious infections (35.9 per 1000 patient-years) was significantly lower than those reported in other post-marketing observational studies, but substantially similar to those reported in phase III RCTs. The most frequent infection site was the lower respiratory tract, with 14 cases of pneumonia (19.2%) and one case of TB. Another four cases of extrapulmonary TB were also reported, two of which occurred during the pre-screening era. No differences among the individual anti-TNF agents were detected. An increased risk of serious infections was associated with age, ESR and corticosteroid use, but not with anti-TNF therapy. Take-home messages • The registries are increasingly being used to investigate the incidence of adverse events in RA patients. • The overall rate of serious infections in our study was significantly lower than those reported in other postmarketing observational studies. • The most frequent infection site was the lower respiratory tract. • No differences among the individual anti-TNF agents were detected. • An increased risk of serious infections was associated with age, ESR and corticosteroid use, but not with anti-TNF therapy.

References [1] Moreland LW, Baumgartner SW, Schiff MH, Tindal EA, Fleischmann RM, Weaver AL, et al. Treatment of rheumatoid arthritis with a recombinant human tumour necrosis factor (p75)-Fc fusion protein. N Engl J Med 1997;337:141–7. [2] Maini R, St Clair EW, Breedveld F, Furst D, Kalden J, Weisman M, et al. Infliximab (chimeric anti-tumour necrosis factor alpha monoclonal antibody) versus placebo in rheumatoid arthritis patients receiving concomitant methotrexate: a randomised phase III trial. Lancet 1999;354:1932–9. [3] Weinblatt ME, Keystone EC, Furst DE, Kavanaugh AF, Chartash EK, Segurado OG. Long term efficacy and safety of adalimumab plus methotrexate in patients with rheumatoid arthritis: ARMADA 4 year extended study. Ann Rheum Dis 2006;65:753–9. [4] Klareskog L, van der Heijde DMFM, de Jager JP, Gough A, Kalden J, Malaise M. Therapeutic effect of the combination of etanercept and methotrexate compared with each treatment alone in patients with rheumatoid arthritis: double-blind randomised controlled trial. Lancet 2004;363:675–81. [5] Lipsky PE, van der Heijde DMFM, St Clair EW, Furst DE, Bredveld FC, Kalden JR, et al. Infliximab and methotrexate in the treatment of rheumatoid arthritis. N Engl J Med 2000;343:1594–602. [6] Keystone EC, Kavanaugh AF, Sharp JT, Tannenbaun H, Hua Y, Teoh LS, et al. Radiographic, clinical, and functional outcomes of treatment with adalimumab (a human anti-tumor necrosis factor monoclonal antibody) in patients with active rheumatoid arthritis receiving concomitant methotrexate therapy: a randomized, placebo-controlled, 52-week trial. Arthritis Rheum 2004;50:1400–11. [7] Baum J. Infections in rheumatoid arthritis. Arthritis Rheum 1971;14:135–1357. [8] Doran MF, Crowson CS, Pond GR, O'Fallon WM, Gabriel SE. Frequency of infection in patients with rheumatoid arthritis compared with controls. A population-based study. Arthritis Rheum 2002;9:2287–93. [9] Ehlers S. Why does tumor necrosis factor targeted therapy reactivate tuberculosis? J Rheumatol 2005;32(Suppl 74):35–9. [10] Gómez-Reino JJ, Carmona L, Valverde VR, Mola EM, Montero MD, for the BIOBADASER Group. Treatment of rheumatoid arthritis with tumor necrosis factor inhibitors may predispose to significant increase in tuberculosis risk: a multicenter active-surveillance report. Arthritis Rheum 2003;48:2122–7. [11] Carmona L, Gomez-Reino JJ, Rodriguez-Valverde V, Montero D, PascualGomez E, Mola EM, et al. Effectiveness of recommendations to prevent reactivation of latent tuberculosis infection in patients treated with tumor necrosis factor antagonists. Arthritis Rheum 2005;52:1766–72. [12] Kalyoncu U, Karadag O, Akdogan A, Kisacik B, Erman M, Erguven S, et al. Pneumocystis carinii pneumonia in a rheumatoid arthritis patient treated with adalimumab. Scand J Infect Dis 2007;39:475–8. [13] Lassoued S, Zabraniecki L, Marin F, Billey T. Toxoplasmic chorioretinitis and antitumor necrosis factor treatment in rheumatoid arthritis. Semin Arthritis Rheum 2007;36:262–3. [14] Jain VV, Evans T, Peterson MW. Reactivation histoplasmosis after treatment with anti-tumor necrosis factor alpha in a patient from a nonendemic area. Respir Med 2006;100:1291–3. [15] Fabre S, Gibert C, Lechiche C, Jorgensen C, Sany J. Primary cutaneous Nocardia otitidiscaviarum infection in a patient with rheumatoid arthritis treated with infliximab. J Rheumatol 2005;32:2432–3. [16] Salliot C, Gossec L, Ruyssen-Witrand A, Luc M, Duclos M, Guignard S, et al. Infections during tumour necrosis factor-alpha blocker therapy for rheumatic diseases in daily practice: a systematic retrospective study of 709 patients. Rheumatology 2007;46:327–34. [17] Listing J, Strangfeld A, Kary S, Rau R, von Hinueber U, Stoyanova-Scholz M, et al. Infections in patients with rheumatoid arthritis treated with biologic agents. Arthritis Rheum 2005;52:3403–12. [18] Dixon WG, Watson K, Lunt K, Hyrich KL, British Society for Rheumatology Biologics Register Control Centre Consortium, Silman AJ, Symmons DPM. Rates of serious infection, including site-specific and bacterial intracellular infection, in rheumatoid arthritis patients receiving antitumor necrosis factor therapy. Arthritis Rheum 2006;54:2368–76. [19] Curtis JR, Patkar N, Xie A, Martin C, Allison JJ, Saag M, et al. Risk of serious bacterial infections among rheumatoid arthritis patients exposed to tumor necrosis factor alpha antagonists. Arthritis Rheum 2007;56:1125–33. [20] Bernatsky S, Hudson M, Suissa S. Anti-rheumatic drug use and risk of serious infections in rheumatoid arthritis. Rheumatology 2007;46: 1157–60. [21] Sarzi-Puttini P, Antivalle M, Marchesoni A, Favalli EG, Gorla R, Fillipini M, et al. Efficacy and safety of anti-TNF agents in the Lombardy Rheumatoid Arthritis Network (LORHEN). Reumatismo in press. [22] Furst DE, Breedveld FC, Kalden JR, Smolen JS, Antoni CE, Bijlsma JWJ, et al. Updated consensus statement on biological agents for the treatment of rheumatoid arthritis and other rheumatic diseases (May 2002). Ann Rheum Dis 2002;61(Suppl 2):2, doi:10.1136/ard.61.suppl_2.ii2.

E.G. Favalli et al. / Autoimmunity Reviews 8 (2009) 266–273 [23] Dixon WG, Symmons DP, Lunt M, Watson KD, Hyrich KL, Silman AJ. Serious infection following anti-tumor necrosis factor alpha therapy in patients with rheumatoid arthritis: lessons from interpreting data from observational studies. Arthritis Rheum 2007;56:2896–904. [24] Kroesen S, Widmer AF, Tyndall A, Hasler P. Serious bacterial infections in patients with rheumatoid arthritis under anti-TNF-α therapy. Rheumatology 2003;42:617–21. [25] Greene C, Lowe G, Taggart C, Gallagher P, McElvaney N, O'Neill S. Tumor necrosis factor-α-converting enzyme: its role in community-acquired pneumonia. J Infect Dis 2002;186:1790–6. [26] Wolfe F, Caplan L, Michaud K. Treatment for rheumatoid arthritis and the risk of hospitalization for pneumonia. Association with prednisone, disease-modifying antirheumatic drugs, and anti-tumor necrosis factor therapy. Arthritis Rheum 2006;54:628–34. [27] Groves RW, Allen MH, Ross EL, Barker JN, MacDonald DM. Tumour necrosis factor α is pro-inflammatory in normal human skin and modulates cutaneous adhesion molecule expression. Br J Dermatol 1995;132:345–52. [28] Khanna D, McMahon M, Furst DE. Safety of tumour necrosis factor-α antagonists. Drug Saf 2004;27:307324. [29] Gardam MA, Keystone EC, Menzies R, Manners S, Skamene E, Long R, et al. Anti-tumour necrosis factor agents and tuberculosis risk: mechanisms of action and clinical management. Lancet Infect Dis 2003;3:148–55.

273

[30] Wallis RS, Broder MS, Wong JY, Hanson ME, Beenhouwer DO. Granulomatous infectious diseases associated with tumor necrosis factor antagonists. Clin Infect Dis 2004;38:1261–5. [31] The World Health Organization; 2005 [http://www.who.int/globalatlas/ dataQuery]. [32] Wallis RS, Ehlers S. Tumor necrosis factor and granuloma biology: explaining the differential infection risk of etanercept and infliximab. Semin Arthritis Rheum 2005;34(Suppl 1):34–8. [33] Fleischmann R, Iqbal I, Nandeshwar P, Quiceno A. Safety and efficacy of disease-modifying anti-rheumatic agents: focus on the benefits and risks of etanercept. Drug Saf 2002;25:173–97. [34] Keane J, Gershon S, Wise RP, Mirable-Levens E, Kasznica J, Schwieterman WD, et al. Tuberculosis associated with infliximab, a tumor necrosis factor alpha-neutralizing agent. N Engl J Med 2001;345:1098–104. [35] Valesini G, Iannuccelli C, Marocchi E, Pascoli L, Scalzi V, Di Franco M. Biological and clinical effects of anti-TNFalpha treatment. Autoimmun Rev 2007;7:35–41. [36] Tincani A, Andreoli L, Bazzani C, Bosiso D, Sozzani S. Inflammatory molecules: a target for treatment of systemic autoimmune diseases. Autoimmun Rev 2007;7:1–7.

HLA-DRB1*0404 is strongly associated with high titers of anti-cyclic citrullinated peptide antibodies in rheumatoid arthritis. It has been widely explored if there is an association between genetic risk factors for rheumatoid arthritis (RA) and the production of anti-CCP antibodies. Several studies have described such relation however the HLA-DRB1*0404 allele has been mainly related with it. In a recent study performed by Charpin C. et al. (Clin Exp Rheumatol 2008; 26: 627-31); they tested if the presence of some alleles such as HLA-DRB1*0101, *0401 and *0404 individually influences the production of anti-CCP antibodies in carriers compared to non-carriers. The frequency of anti-CCP antibodies was calculated in the sera of 260 RA patients expressing either two (double dose genotypes SE+/SE+), one (single dose genotypes SE+/SE-) or no RA associated HLA-DR alleles (SE-/SE-). Anti-CCP antibodies titers were also determined. As result, they found that the presence of HLA-DR alleles is not mandatory for production of anti-CCP antibodies. They found that 68% of SE-/SE- patients were anti-CCP positive. There was no significant difference in the presence of anti-CCP antibodies between SE negative patient (SE-/SE-) and patients expressing at least one SE (SE+/SE+ and SE+/SE-) (p=0.140). They observed no statistical difference in anti-CCP between RA patients expressing one or two SE (82% vs. 77%, p=0.577) as well. Among SE+/SE-patients, HLA-DRB1*0404 was associated with anti-CCP with a statistically significant difference after compared with SE negative patients (90% anti-CCP positive, p=0.02). HLA-DRB1*0404 was also associated with high titers of anti-CCP antibodies with a statistically significant difference after compared with HLA-DRB1*0401 and HLA-DRB1*0101 patients (p=0.025). They finally concluded that RA-associated HLADRB1*0404 allele was the most strongly allele associated with the presence of anti-CCP antibodies in RA sera. Moreover, HLADRB1*0404 patients had higher titers of anti-CCP than patients with other RA associated HLA-DR alleles, confirming previous findings on this subject.

Anti-DNA Ig peptides promote Treg cell activity in systemic lupus erythematosus. In murine lupus models, treatment with anti-DNA Ig-based peptides can expand the number of Treg cells in vivo. This study was undertaken to test the possibility that functional Treg cells can be induced by exposure to ant-DNA Ig-based peptides. Hahn B. et al (Arthritis Rheum 2008; 58: 2488-97). Peripheral blood mononuclear cells were isolated from 36 lupus patients and 32 healthy individuals. Short-term culture experiments in the presence of several independent stimuli including anti-DNA Ig peptides were followed by flow cytometric analysis for identification of CD4+CD25+ T cells, cell sorting for in vitro suppression assays, and analysis of correlations between the expansion of FoxP3 and serologic and clinical characteristics of the SLE patients. The number of in vitro CD4+CD25+ T cells increased after culture with anti-DNA Ig peptides in the SLE patients, but not in the controls. The expanded CD4+CD25+ T cells required FoxP3 for cell contact-mediated suppression of proliferation and interferon-gamma production in target CD4+CD25- t cells. The induction of FoxP3 in SLE Treg cells occurred only in seropositive patients, and was correlated with anti-DNA and IgG serum titers. These results suggest a new modality to reverse the functional deficit of Treg cells in SLE patients with positive autoimmune serology, and identify a new strategy to enhance immunoregulatory T cell activity in human SLE.