Abstract
Objective. To assess longterm safety in a US cohort of patients with rheumatoid arthritis (RA) treated with adalimumab (ADA) in real-world clinical care settings.
Methods. This observational study analyzed the longterm incidence of safety outcomes among patients with RA initiating ADA, using data from the Corrona RA registry. Patients were adults (≥ 18 yrs) who initiated ADA treatment between January 2008 and June 2017, and who had at least 1 followup visit.
Results. In total, 2798 ADA initiators were available for analysis, with a mean age of 54.5 years, 77% female, and mean disease duration of 8.3 years. Nearly half (48%) were biologic-naive, and 9% were using prednisone ≥ 10 mg at ADA initiation. The incidence rates per 100 person-years for serious infections, congestive heart failure requiring hospitalization, malignancy (excluding nonmelanoma skin cancer), and all-cause mortality were 1.86, 0.15, 0.64, and 0.33, respectively. The incidence of serious infections was higher in the first year of therapy (3.44, 95% CI 2.45–4.84) than in subsequent years, while other measured adverse effects did not vary substantially by duration of exposure. The median time to ADA discontinuation was 11 months, while the median time to first serious infection among those experiencing a serious infection event was 12 months.
Conclusion. Analysis of longterm data from this prospective real-world registry demonstrated a safety profile consistent with previous studies in patients with RA. This analysis did not identify any new safety signals associated with ADA treatment and provides guidance for physicians prescribing ADA for extended periods.
Rheumatoid arthritis (RA) is a complex immune-mediated inflammatory disease, which, if not treated effectively, can cause significant pain and progressive joint damage leading to disability and a reduced quality of life1,2. When compared with the general population, RA is associated with increased morbidity due to complications and comorbidities such as serious infections, cardiovascular (CV) disease, and certain cancers3,4,5. Higher mortality rates have also been reported3,5. Inhibition of tumor necrosis factor (TNF), a key component of the inflammatory mechanism associated with many immune-mediated diseases, has been shown to reduce RA disease activity and improve clinical, radiographic, and functional outcomes6,7.
TNF inhibitors (TNFi) have been associated with a reduced risk of CV disease in patients with RA8,9; however, some uncertainty remains on the increased risk of infections and the increased potential for malignancy and other adverse events (AE) in patients treated with TNFi. Prior studies have reported an increased risk of serious infections and malignancies, particularly lymphoma, in patients with psoriasis treated with TNFi, while other more recent studies have not identified similar significantly increased risks among patients treated with TNFi10–18. Because TNF is an important component of the human immune system and has a role in tumor growth mediation, the link between its inhibition and an increase in the frequency of serious infections or malignancies deserves more investigation19.
To date, more than 20 years after the introduction of TNFi, the safety of biologics such as adalimumab (ADA) has been established through both randomized controlled clinical trials (RCT) and registry-based European observational studies17,18,20,21,22. A large cross-indication (including RA, juvenile idiopathic arthritis, ankylosing spondylitis, psoriatic arthritis, psoriasis, and Crohn disease) analysis of longterm safety based on almost 30,000 patients treated with ADA in global clinical trials and open-label extension studies did not identify any new safety signals, although individual differences between the disease populations did exist17,18. Additionally, in these studies the mortality rate for most indications was below what would be expected in an age- and sex-adjusted population17,18. However, the patient populations from RCT and open-label extension studies may not be reflective of a real-world population of patients with RA in the United States. Comparisons between different registry-based studies are also challenging because of differences in design, followup assessments, and other variables assessed within each registry, which may result in heterogeneity between studies23. For example, the percentage of patients with a high Disease Activity Score has been shown to vary between registries, with patients in the Corrona registry having lower disease activity and higher functional status23,24. Differences in healthcare delivery systems and baseline patient characteristics between European and US populations may result in differences in the underlying risk of certain AE, highlighting the need to establish the safety profile of TNFi in both of these populations.
The longterm incidence of AE among ADA users in real-world clinical settings has not been well documented in the US. In a recent US study, about half of biologic-naive patients with RA stated that their unwillingness to initiate treatment with a biologic was primarily due to fear/concern about side effects, highlighting the need to clarify the safety of biologics such as ADA in US real-world practice25. The objective of our study was therefore to build on the existing knowledge base on the safety of TNFi through the evaluation of the risk of AE in relation to duration of ADA exposure in a US-based population. This will complement ongoing longterm clinical trial safety analyses and further test the hypothesis that the real-world safety profile of ADA is consistent with that previously demonstrated by RCT analyses as well as ex-US RA registries.
MATERIALS AND METHODS
Study design
This was a retrospective analysis of the US Corrona RA Registry, a prospective observational cohort of patients with RA initiated in 200126, which contains data on nearly 50,000 patients across 40 states, covering about 373,000 visits and 173,000 person-years (PY) of followup. Data are collected from both patients and their treating rheumatologists, who gather information on disease duration, prognosis, disease severity and activity, comorbidities, use of medications, and patient-reported outcome data. Participating rheumatologists also actively assess for the occurrence of AE at each registry followup visit. Followup assessments are requested as often as every 6 months and completed during routine clinical encounters.
Study population
Patients were included in this analysis if they were adults (≥ 18 yrs of age), initiated ADA as therapy for a clinical diagnosis of RA between January 1, 2008, and June 1, 2017, and were enrolled in the US Corrona RA Registry. Patients were also required to have at least 1 followup visit post-drug initiation. Patients with RA onset prior to 16 years of age were excluded from the study.
Ethics
All patients provided written informed consent prior to registry participation. Institutional review board (IRB) approvals for this study were obtained from a central IRB (New England Independent Review Board, NEIRB No. 02-021) for private practice sites and local IRB of participating academic sites.
Exposure
The date on which each patient initiated ADA therapy was defined as their index date and served as the starting point for the measurement of both the time to first AE and the incidence of the AE over time.
Outcomes and covariates
The safety outcomes assessed in this analysis were physician-reported serious infections (i.e., those requiring hospitalization or intravenous antibiotics), malignancies (excluding nonmelanoma skin cancer), congestive heart failure (CHF) that required hospitalization, tuberculosis (TB) with a focus on active TB, and drug-induced systemic lupus erythematosus. All-cause mortality was also assessed. Outcomes were assessed by both time from ADA initiation to the first occurrence of a specific AE, and the frequency, or incidence, of each specific AE over time. Measurement of these outcomes continued until either the first occurrence of the specific AE, or for incidence of the specific AE over time, until 90 days following discontinuation of ADA or the patient’s last Corrona visit (whichever occurred first), at which point the patient was censored. For malignancy and mortality, all events occurring during followup were included regardless of ADA continuation or discontinuation.
The Corrona registry uses an established system for the validation of physician-reported AE. In brief, treating physicians complete Targeted Adverse Event questionnaires to record serious AE. These questionnaires, alongside supporting documents appropriate to the event (e.g., hospitalization records, pathology reports), are submitted to Corrona for validation. A subset of questionnaires is also studied for expert adjudication. This methodology has been supported by previous validation studies, which found positive predictive values of Targeted Adverse Event forms to be 96% for CV events27, 86% for malignancies28, and 71% for serious infectious events29.
Analysis
Demographic and disease characteristics, including prior treatment history, of all patients at the time of their ADA therapy initiation were recorded. Percentages were used to describe categorical variables while means and SD were used to describe continuous variables. In cases of highly skewed distributions, medians and interquartile ranges were used to describe continuous variables.
The time from ADA initiation to the first occurrence of a specific AE was included in the analysis of incident events. For each specific AE, only patients without a history of that AE prior to or at the time of ADA initiation were followed for the occurrence of the AE.
Incidence rates (IR) and 95% CI are reported per 100 PY, assuming a Poisson distribution. Time to first event for particular events, including time to discontinuation, time to serious infection, and time to malignancy, was evaluated using Kaplan-Meier curves. Median (95% CI) time to first event was calculated as well as the proportion of patients who did not have an event at 6, 12, 24, 36, 48, 60, 72, and 84 months after ADA initiation. The standardized mortality ratio (SMR) was calculated based on recent Centers for Disease Control and Prevention mortality data30.
RESULTS
Study population at ADA initiation
In total, 2798 real-world patients who had initiated ADA therapy were available for analysis. Patient, disease, and prior treatment characteristics of these patients at time of ADA initiation are presented in Table 1. The mean (SD) age of these patients, who were predominantly female (77%), was 54.5 years (12.3 yrs), with a mean (SD) body mass index of 30.4 (7.5). About half of the patients (52%) had never smoked. The mean (SD) duration of disease was 8.3 (9) years, and the mean Clinical Disease Activity Index score was 20.4 (14), with 39% of the patients in the high disease activity category. About half (48%) were biologic-naive, and 9% were treated with ≥ 10 mg prednisone daily. The majority of patients (60%) were receiving concomitant methotrexate at ADA initiation. Additional patient, disease, and treatment characteristics are in Supplementary Table 1 (available with the online version of this article).
Patient, disease, and prior treatment characteristics at ADA initiation.
Duration of ADA exposure during the assessment period
Overall, the majority of patients (84%) received ADA therapy for < 3 years (Table 1). The median time from ADA initiation to last visit on ADA therapy was 11 months (95% CI 4–25). The proportion of patients remaining on ADA therapy at 6 and 12 months after initiation was 70% and 56%, respectively. In the subset of patients who were biologic-naive at ADA initiation, the median time from ADA initiation to last visit on ADA therapy was 13 months (95% CI 6–29), with 78% and 62% remaining on ADA at 6 and 12 months, respectively (data not shown).
The probability of patients within the analysis remaining on ADA over time is presented as a time to discontinuation from ADA in Figure 1. The 3 most commonly reported reasons for discontinuing ADA treatment among the overall population included failure to maintain initial response (27%), inadequate initial response (16%), and minor side effect (14%); serious side effects accounted for 6% of reported reasons for discontinuation (data not shown). Similar frequencies of reasons for discontinuation were observed among the biologic-naive population.
Probability of continuing adalimumab treatment.
Incidence of events during ADA exposure
The IR per 100 PY of ADA exposure was 1.86 (95% CI 1.50–2.31) for serious infections, 0.15 (0.07–0.31) for CHF requiring hospitalization, 0.64 (0.50–0.84) for malignancy (excluding nonmelanoma skin cancers), and 0.33 (0.24–0.48) for all-cause mortality (Table 2). No cases of TB were reported during followup. The IR per 100 PY of ADA exposure for AE included in this study (excluding malignancy and mortality) leading to ADA discontinuation was 0.73 (0.52–1.03).
Incidence of AEa among ADA initiators.
The incidence of events by duration of ADA exposure categories (≤ 1 yr, > 1 to 3 yrs, > 3 to 5 yrs, > 5 yrs) is presented in Figure 2. The incidence of serious infections was higher in the first year of therapy (3.44 per 100 PY, 95% CI 2.45–4.84) than in subsequent years, while other AE did not substantially vary by duration of exposure.
Incidence of events by duration of adalimumab exposure. Error bars indicate 95% CI. Serious infections included those that led to hospitalization or IV antibiotics. Malignancy excluded nonmelanoma skin cancer. CHF: congestive heart failure; IV: intravenous.
Time to first occurrence of AE of interest
Median times to first serious infection and malignancy among those experiencing each respective event were 12 months (95% CI 7–26) and 34 months (16–56), respectively. The proportion of patients not having a first occurrence of serious infection and malignancy at 6 months was 98.4% and 99.7%, respectively. This remained largely unchanged at 12 months, at which 97.5% and 99.4% of patients had not had an occurrence of serious infection or malignancy, respectively. Throughout the remainder of the evaluation period, the large majority of patients did not have an occurrence of serious infection or malignancy at 24, 36, 48, 60, 72 or 84 months (Supplementary Table 2, available with the online version of this article). Kaplan-Meier curves for the probability of not having a serious infection or having a malignancy event are presented in Figure 3 and Figure 4, respectively.
Probability of not having a serious infection.
Probability of not having a malignancy event.
SMR data
The age-adjusted and sex-adjusted SMR comparing the observed and expected number of deaths was 0.50 ± 0.32. The SMR reveals that there were an expected 62 mortalities within the ADA-initiating population, compared to only 31 deaths observed (data not shown).
DISCUSSION
With an increasing number of physicians prescribing disease-modifying antirheumatic medications for longterm chronic use, knowledge of the longterm safety profiles of such medications is critical in guiding providers in prescribing strategies that minimize the risk of serious AE for their patients. This retrospective analysis of a real-world US database of patients with RA treated with ADA immunotherapy is consistent with the safety profile previously established through RCT, open-label extension studies, and European RA registries.
An update of global longterm clinical trial safety data in adult patients treated with ADA for multiple indications (predominately RA studies), which included 29,987 patients with 56,951 PY of exposure, identified serious infections as the most frequent AE, with an IR of 4.6 events per 100 PY18. Our study found a lower IR per 100 PY for serious infections, which was higher during the first year of exposure (3.44, 95% CI 2.45–4.84) than in subsequent years. Previous studies have also shown the risk of infection to be higher earlier in treatment compared with later in treatment31. It may be the case that these AE were more strictly reported in the clinical trials, as compared with the Corrona registry, although given the serious nature of these AE, this seems unlikely. Patients in Corrona may have lower levels of disease activity compared to patients in the RCT. Previous studies have shown that increased RA disease activity is associated with an increased risk for serious infections24,32,33. Differences in the methodology used for the calculation of AE IR may also help explain the differences in the observed rates. Compared to the present study, which collected AE at the patient level, the global longterm clinical trials study collected AE at the event level, with multiple AE of the same type within an individual reflected in the IR. The global clinical trial analysis reported a lower number of deaths in patients with RA treated with ADA (SMR of 0.74) than would be expected in an age- and sex- adjusted population18, a number directionally similar to the SMR observed in our real-world analysis. For both analyses, underlying explanations for this result are not immediately clear.
Comparison of the Corrona registry findings with those from other real-world databases also reveals variations in outcomes across these analyses. The British Society for Rheumatology Biologics Register for Rheumatoid Arthritis (BSRBR-RA) reported that no TNFi was associated with a statistically higher risk of serious infection than any other TNFi among a total of 19,282 patients contributing 46,771 PY of followup. BSRBR-RA reported an IR for serious infection of 5.42 per 100 PY among 7818 patients initiating ADA, which was higher than the first-year IR of serious infection reported for ADA-exposed patients in our study (IR 3.44)22. The IR for malignancy during a median followup of 3.5 years per patient in a BSRBR-RA cohort exposed to TNFi was more similar, although still slightly higher, than that observed in the present study (0.77 vs 0.64, respectively)21. During a total of 3 years of followup, the Anti-Rheumatic Therapy in Sweden biologics register reported an IR for mortality of 1.4 per 100 PY after exposure to ADA. In both studies, the IR observed may be higher than that reported in our analysis because of differences in median followup time20. Additionally, patients in Europe often have more restricted access to biologics and thus have higher levels of disease activity, greater use of prednisone, and higher levels of disability prior to biologic initiation, all of which influence the risk of infection24,34. Prevalence of clinical characteristics (e.g., smoking) and comorbidities (e.g., obesity) also varies between US and European registries, and may reflect the way comorbidity data were recorded as well as general cultural differences between countries. Differences in comorbidity profiles could create differences in the IR of AE23,24.
Comparing safety outcomes for ADA in US clinical practice using data from a large, nationwide cohort of patients with RA is a vital strength of our study. Owing to payer and regulatory differences, patient characteristics and access to biologic drugs can vary substantially between countries. Therefore, it is essential to obtain representative clinical evidence directly relevant to practicing rheumatologists in the US. A previous study showed that Medicare beneficiaries who were enrolled in Corrona share similar demographic and clinical characteristics with Medicare beneficiaries with claims for RA or visits to a rheumatology specialist who were not enrolled, thus suggesting that the results presented here may be generalizable to the broader RA population in the US35. The relative similarity of the AE findings of our analysis compared with prior clinical trials and other observational registry studies is reassuring, yet also highlights the importance of continuing to assess these outcomes over longer periods of time, and in more diverse populations.
Our study also has several limitations. First, ADA exposure time was limited for some patients. In total, 50% of patients used ADA for 1 year or less and only 16% of patients were exposed for 3 years or more. While serious infection event rates tend to be highest during the first 6–12 months of TNFi exposure36,37, the limited number of patients in our study with longer-term ADA exposure may not have allowed us to fully characterize the rates of serious infection as well as other AE such as malignancies because of the relative infrequency and longer-term onset of these AE. There is an inability to determine with certainty whether ADA exposure had an effect on the development of the AE measured, particularly those AE occurring after drug discontinuation, or conditions that typically take many years to develop (e.g., malignancy). We did not account for subsequent therapies initiated after ADA discontinuation and their potential effect on the occurrence of AE observed in our study. The SMR among this population was low, which could have caused undercounting of events within the Corrona registry. However, serious events are less likely to be underreported, and a low SMR was also found in the ESPRIT study, which evaluated the longterm safety and effectiveness of ADA in the treatment of psoriasis, and in studies evaluating the longterm safety of ADA across multiple indications15,17,18. Finally, the precision with which ADA discontinuation is measured in the Corrona registry may have resulted in misclassified exposure time. The date of discontinuation was recorded at the time of physician visit; however, there may be challenges with recall bias if the discontinuation is not well documented in the physician’s clinical notes.
This analysis of targeted data from the prospective Corrona real-world RA registry demonstrated ADA to have a safety profile consistent with previous studies and did not identify any new safety concerns with longterm ADA treatment. Therefore, with the addition of these real-world data in a large sample of US patients with RA, the previously understood benefit-risk profile of ADA for the treatment of RA remains unchanged. This confirmation of existing knowledge may reassure providers who are initiating and monitoring ADA in their patients.
Acknowledgment
Medical writing support was provided by Brandy Menges of JK Associates Inc., and editorial support was provided by Sebastian Reynolds and Fiona Woodward of JK Associates Inc. This support was funded by AbbVie.
Footnotes
Full Release Article. For details see Reprints and Permissions at jrheum.org.
Financial support for the study was provided by AbbVie, which manufactures adalimumab. AbbVie participated in the interpretation of data, and review and approval of the manuscript. The study was sponsored by Corrona LLC. Corrona LLC has been supported through contracted subscriptions in the last 2 years by AbbVie, Amgen, Boehringer Ingelheim, Bristol-Myers Squibb, Crescendo, Eli Lilly and Co., Genentech, Gilead, GSK, Horizon Pharma USA, Janssen, Momenta Pharmaceuticals, Novartis, Pfizer Inc., Roche, Merck, UCB, and Valeant. The study design and conduct were a collaboration between Corrona and AbbVie. L.R. Harrold is an employee and shareholder of Corrona LLC. C.J. Barr is an employee and shareholder of Corrona LLC. H.J. Litman is an employee of Corrona LLC. J. Griffith, S.S. Islam, D. Guo, J. Fay, and P. Zueger are employees of AbbVie Inc. and own AbbVie stock. J.D. Greenberg is an employee and shareholder of Corrona LLC. L.R. Harrold, MD, MPH, University of Massachusetts Medical School, and Corrona LLC; J. Griffith, PharmD, AbbVie; P. Zueger, PharmD, PhD, AbbVie; H.J. Litman, PhD, Corrona LLC; B. Gershenson, MPH, University of Massachusetts Medical School; S.S. Islam, MD, MSPH, DrPH, AbbVie; C.J. Barr, MPH, Corrona LLC; D. Guo, MD, PhD, AbbVie; J. Fay, MD, AbbVie; J.D. Greenberg, MD, MPH, Corrona LLC, and New York University School of Medicine.
- Accepted for publication July 29, 2019.
Free online via JRheum Full Release option
REFERENCES
ONLINE SUPPLEMENT
Supplementary material accompanies the online version of this article.
