Get Connected:


Comparative Effectiveness of TNF Blockers in Rheumatoid Arthritis Patients

Among rheumatoid arthritis patients newly treated with adalimumab, etanercept, or infliximab, lower dose escalation rates and costs were seen with etanercept.
Published Online: Dec 04,2013
Benjamin Chastek, MS; Stuart D. Segal, MD; R. Peter Bonafede, MD; Crystal Watson, MS; Laura Becker, MS; Sandeep Chaudhari, MS; David J. Harrison, PhD; and Shravanthi R. Gandra, PhD, MBA
Objective: To estimate dose escalation patterns among rheumatoid arthritis (RA) patients newly treated with adalimumab, etanercept, or infl iximab in clinical practices, and the association of dose escalation with clinical outcomes and costs.

Study Design: Retrospective cohort study using claims data and linked medical records.

Methods: Dosing information was evaluated for RA patients initiating treatment with adalimumab, etanercept, or infl iximab from January 1, 2006, to May 31, 2008, and continuing treatment for 1 year; outcomes include percentages of patients with dose escalation and total costs for tumor necrosis factor (TNF) blockers. Clinical information was reviewed by rheumatologists, who assigned clinical change scores from baseline until the visit closest to 1 year after index date. Logistic regressions, adjusted for baseline patient characteristics, compared dose escalation rates and clinical responses among medication cohorts. Costs were analyzed descriptively using t tests to examine differences between cohorts.

Results: Of 1495 patients who met selection criteria, 334 had scored medical records. Etanercept had the lowest percentage of patients with dose escalation (1.7% vs 9.5% of adalimumab patients [<.001] and 50.3% of infl iximab patients [P <.001]). Among etanercept patients with scored medical records, 85% were rated as “much better/better” versus 81% of adalimumab patients (P = .387) and 77% of infl iximab patients (P = .111). Etanercept patients incurred lower mean TNF blocker costs ($21,821) than adalimumab ($23,117; P <.001) and infl iximab ($23,561; P = .001) patients.

Conclusions: Etanercept was associated with lower dose escalation rates, similar clinical change scores, and lower TNF blocker costs than adalimumab and infliximab.

Am J Pharm Benefits. 2013;5(Special Issue):SP14-SP22
Rheumatoid arthritis (RA), an autoimmune disease affecting approximately 1% of the US population,1 is characterized by chronic, persistent synovial inflammation, cartilage loss, bone erosions in peripheral joints,2 and a wide array of multisystem comorbidities.3 Early, effective therapy is needed to manage this chronic progressive disease.3 For moderate to severe RA patients, 3 of the most common tumor necrosis factor (TNF) blockers (adalimumab, etanercept, and infliximab) have been shown to be effective in managing this disease,4-7 with response rates of up to 70% in treatment groups of randomized clinical trials.8

Etanercept is a recombinant human soluble TNF-receptor protein; adalimumab and infliximab are anti-TNF monoclonal antibodies. Doses of these medications recommended by the US Food and Drug Administration for the treatment of RA are adalimumab 40 mg every other week, with the option of increasing to 40 mg/week9; etanercept 50 mg/week10; and infl iximab 3 mg/kg at weeks 0, 2, and 6 and then every 8 weeks thereafter, with the option of increasing the dose to 10 mg/kg or the frequency to every 4 weeks.11

Numerous US and European studies have shown higher dose escalation rates among patients treated with infliximab and adalimumab than among those treated with etanercept.8,12-21 However, no published US studies have taken clinical outcomes into account while examining dosing patterns and costs of TNF blockers in clinical practice. Dose escalation is associated with higher drug treatment costs13,14,22 and greater risk for adverse events such as infections and infusion reactions,23-25 but the clinical benefit of dose escalation in treatment groups may be limited12 or unsustained.8

By evaluating the association of dose escalation with clinical outcomes and TNF blocker costs among patients with moderate to severe RA treated with adalimumab, etanercept, or infliximab in community-based US practice, this study addresses a gap in comparative effectiveness research, potentially providing useful information to US payers.


Study Design and Data Sources

This study was a retrospective claims analysis designed to evaluate utilization, dosing, and cost information for eligible patients and a medical record review to evaluate clinical outcomes for a population subset. Claims and enrollment data were obtained from a large, nationally representative US-managed healthcare plan affiliated with OptumInsight. In 2009, this plan covered about 13 million patients with medical and pharmacy benefits.

For a subpopulation, medical records linked to claims data (which included physician notes, radiographic reports, laboratory reports, and patient-completed assessments) were obtained from providers. These were reviewed by independent rheumatologists to identify patients’ clinical outcomes. Physician participation in providing medical record data was voluntary. Medical records were accessed in accordance with protocols compliant with the Health Insurance Portability and Accountability Act. Following institutional review board and health plan approval, de-identified medical records for eligible patients were obtained by a firm specializing in medical
record abstraction.

Practicing rheumatologists developed an algorithm to determine patients’ clinical response, or clinical change score, based on the criteria used by the American College of Rheumatology,26 the World Health Organization/International League of Associations for Rheumatology,27 and disease activity measures commonly used in clinical trials. The rheumatologists assigned a score to each chart based on changes from baseline (visit before or on the day treatment started) in 6 key RA-related outcomes: patient’s joint pain, patient’s global assessment, synovitis, other (morning stiffness, fatigue, physician’s global assessment, or radiographic reports), laboratory values, and functional assessment measures that were available in the chart. Patients had to have at least 1 key RA-related  outcome over time that could be scored; missing measureswere considered equivalent to “no change.” Each de-identified medical record was reviewed and scored by 2 independent pairs of rheumatologists from different geographic areas. Each pair assigned a clinical change score based on a 5-category response scale: “much better,” “better,” “no change,” “worse,” and “much worse.” The final clinical change score was derived by overall clinical judgment based on the key RA-related outcomes available in the medical records. If the rheumatologist  pairs assigned different overall clinical change scores, they discussed them and agreed upon a final score. If they were unable to reach agreement, a third pair of rheumatologists reviewed the medical record and determined the final score.

Patient Identification

Biologic-naïve RA patients initiating treatment with adalimumab, etanercept, or infl iximab between January 1, 2006, and May 31, 2008, were identifi ed from claims. The date of the fi rst fi ll/infusion for the study medication was defined as the index date. To be included in the study, patients had to meet these criteria: (1) RA diagnosis (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] code 714.xx) in any position on medical claims during the preindex period, defi ned as 6 months preceding the index date; (2) continuous treatment (defi ned as suffi cient fi lls or administrations of the index TNF blocker such that patients had been on treatment for at least 60% of their postindex period and had at least 1 prescription or infusion for the same TNF blocker in 3 of the 4 quarters) with the TNF blocker on the index claim for 12 months following the index date (postindex period); (3) age between 18 and 64 years as of the index date (cost data among patients aged more than 65 years are likely to be incomplete due to coordination of benefi ts with Medicare; therefore, these patients were excluded); and (4) continuous plan enrollment with medical and pharmacy benefi ts for 6 months before and 12 months after the index date.

Patients were excluded if they met any of these criteria: (1) had any biologic agent indicated for RA (Appendix A) during the preindex period; (2) had a diagnosis code for psoriasis (ICD-9-CM code 696.0), psoriatic arthritis (code 696.1), ankylosing spondylitis (code 720.0), Crohn’s disease (code 555.xx), ulcerative colitis (code 556.x), human immunodeficiency virus/acquired immunodeficiency syndrome (codes 042xx, V08xx), or cancer (codes 140xx- 172xx, 174xx-208xx) in any position at any time during the study; (3) received any investigational drug or another nonindex TNF blocker or participated in an interventional trial at any time during the study; or (4) for adalimumab and etanercept patients, did not take their first 2 fills as prescribed (ie, gap between fi lls was 30% longer than days of supply); for infl iximab patients, did not receive their maintenance infusion as recommended (ie, gap between the third and fourth infusions was more than 73 days). For criterion 4, the percentage was identifi ed to correspond with the 30% increase required by the definition for a dose escalation, and 73 days is 30% more than 56 days, which is the expected time between infusions.

Patient Selection for Medical Record Abstraction

Medical records of patients who met selection criteria and whose physicians were willing to provide medical records were included in the fi nal medical record abstraction study subpopulation. Patients had to have a visit to a rheumatologist or TNF blocker prescriber within 1 year ± 3 months after the index date; rheumatologists were targeted since they were considered most likely to record RA outcomes.

Observation Periods

The observation period was 12 months after patients’ index date for claims-based outcomes (dose escalation and cost). Patients in the medical record abstraction subpopulation had their clinical assessment identifi ed as the visit closest to 1 year postindex (ie, 9-15 months postindex), since patients may not have visited their physician exactly 1 year after initiating treatment.

Study Outcomes

The primary study outcome was the proportion of RA patients in each medication cohort with at least 1 postindex escalation. For patients receiving adalimumab or etanercept, the maintenance dose was defi ned as the mean of the average daily doses (ADDs) for the first 2 fills. The ADD was calculated as total number of milligrams received in the first 2 fills divided by number of days from the first until the third fill. For patients receiving infl iximab, the maintenance dose was defi ned as the ADD of the third infusion, calculated as total number of milligrams received on the third infusion divided by number of days between the third and fourth infusions.

One-year dose escalation rates were compared among the cohorts using 3 definitions:
  • Two or more instances: 2 or more consecutive claims after the maintenance dose, each with an ADD that was at least 30% greater than the maintenance dose.
  • Average dose: average of all ADDs after themaintenance dose was at least 30% greater than the maintenance dose.
  • Higher than minimum labeled dose: 2 or more consecutive claims with an ADD that was at least 30% greater than the minimum labeled ADD. The minimum labeled dose was 2.86 mg/day for adalimumab and 7.14 mg/day for etanercept. Because infliximab dosing is weight based and claims data do not contain patient weight, the minimum labeled ADD for patients taking infliximab was determined as total number of milligrams received in the first infusion divided by 56 days. Infliximab escalation patterns were examined starting with the third infusion.
Additional outcomes include 1-year TNF blocker costs rom claims and clinical change scores from medical records, both stratified by medication cohort.

Analytic Strategy

Patient characteristics were analyzed descriptively. Differences were examined using the χ2 test for categorical variables and analysis of variance for continuous variables. All patient characteristics were compared between patients with and without medical record data to assess the extent to which the medical record sample was representative of the claims sample.

Dose Escalation Analysis

Comparisons of percentages of patients with dose escalation were conducted between each pair of cohorts using a χ2 test. The impact of treatment group (etanercept as reference group) on dose escalation was analyzed with a logistic regression model. All logistic regression analyses were adjusted for age; sex; comorbidity score; region of plan enrollment; baseline nonsteroid anti-infl ammatory drug, corticosteroid, and disease-modifying antirheumatic drug use; and index provider specialty. Dose escalation sensitivity analyses were performed by raising the 30% threshold to a 40% higher ADD compared with the maintenance dose or minimum labeled dosing.

Clinical Response Analysis

Clinical response was analyzed by treatment cohort using the overall clinical change score from baseline to the 1-year visit as determined by rheumatologists. Clinical change score categories were collapsed into percentages of patients who achieved “much better/better” versus “no change/worse/much worse.” Rheumatologists viewed “no change” as a worsening of the patient’s condition, since the patient was receiving treatment that could potentially cause side effects without conferring any benefit. Statistical significance was tested using a χ2 test and logistic regression analysis.

Cost Analysis

Total TNF blocker costs were calculated for each patient by multiplying the observed dose by the 2011 September wholesale acquisition cost ($896.35 for adalimumab 40 mg, $452.44 for etanercept 50 mg, $710.12 for infliximab 100 mg) and were analyzed descriptively by cohort. Based on claims data, intravenous administration costs for infl iximab were based on the assumption of 2 hours per infusion. Using the June 2011 Medicare Physician Fee Schedule, infusion fees for the initial hour ($146.44) and 1 subsequent hour ($31.26) were applied to each infl iximab infusion and added to infl iximab drug costs. Within drug cohorts, comparisons were made between patients who had a dose escalation and those who did not.


Sample Selection and Attrition

The fi nal sample size was 1495 patients; of those, 1161 were in the claims-only population (claims data group) and 334 had claims data and a scored medical record with sufficient data to examine overall TNF blocker clinical change score (medical record data group) (Table 1; Appendix B). Adalimumab users represented 31.8% of the patients (n = 476); etanercept users, 44.7% (n = 668); and infliximab users, 23.5% (n = 351).

Patient Demographic and Clinical Characteristics

The mean (standard deviation [SD]) ages of the claims data and scored medical record groups were 48.2 (10.6) and 50.0 (9.6) years, respectively (P = .003) (Table 1). The infl iximab cohort had a signifi cantly higher mean (SD) age (50.1 [9.6] years) than the adalimumab cohort (48.6 [10.2] years; P = .038) and the etanercept cohort (47.8 [10.9] years; P <.001) (Appendix C).

Dose Escalation

Across all definitions, etanercept users were significantly less likely to have a dose escalation compared with adalimumab and infliximab users (P <.05) (Figure 1, Appendix D). Using the defi nition of the greater-than-minimum labeled dose, 1.7% of etanercept patients had a dose escalation versus 9.5% of adalimumab patients (P <.001) and 50.3% of infl iximab patients (P <.001). Sensitivity analysis using the 40% threshold revealed similar results, with 0.5% of etanercept, 8.2% of adalimumab (P <.001), and 40.0% of infliximab (P <.001) patients having a dose escalation.

In logistic regression analysis, using the definition of the greater-than-minimum labeled dose, patients treated with adalimumab (odds ratio [OR] = 6.440; P <.001) or infl iximab (OR = 66.385; P <.001) were more likely to experience a dose escalation than patients treated with etanercept (Table 2). Baseline disease-modifying antirheumatic drug use was also significantly associated with less likelihood of a dose escalation (OR = 0.606; P = .034). No other independent variables were associated with statistically significant differences.

Clinical Response

There were no significant differences in clinical change scores among treatment groups 1 year after TNF blocker initiation. (For measurement of the clinical change score, 1 year actually varied by patient from 9 to 15 months.) Among etanercept patients with a scored medical record (n = 134), 85.1% were “much better/better” (95% confidence interval [CI], 77.9-90.6) versus 80.9% of adalimumab patients (n = 110; 95% CI, 72.3-87.8; P = .387) and 76.7% of infl iximab patients (n = 90; 95% CI, 66.6-84.9; P = .111). Figure 2 shows no signifi cant differences in the percentages of dose escalators versus non–dose escalators who experienced improvement. Although the difference was not statistically signifi cant, etanercept patients had the highest response rate.

Improvements in clinical change score among patients with medical record data, as determined by logistic regression analysis, are shown in Table 2. Adalimumab patients (OR = 0.635; P = .212) and infl iximab patients (OR = 0.509; P = .101) had similar odds of being classified as “better/much better” compared with etanercept patients. No other variables were signifi cantly associated with improvement in clinical response, including any dose escalation (OR = 0.997; P = .994).

Cost Analysis

Etanercept dose escalators had 2% higher TNF blocker costs compared with non–dose escalators (P = .027); adalimumab escalators had 35% higher costs compared with non–dose escalators (P <.001); and infl iximab escalators had 44% higher costs compared with non–dose escalators (P <.001) (Figure 3). Infl iximab patients (n = 350) had a mean (SD]) infusion count of 8.3 (1.5), with an average infliximab administration cost of $1485 ($267) per patient, which was 6.7% of the total infliximab drug cost.


The combined administrative claims and medical record data in this study provide detailed information on dosing patterns in the routine clinical setting, as well as on associated clinical outcomes and treatment costs. Such data provide key information for payers making formulary decisions. This study is unique in that it contains both clinical outcomes data and actual drug utilization data from US clinical practices across 3 RA treatment cohorts. The key finding of this study is that although patients on etanercept were significantly less likely to have a dose escalation compared with patients on adalimumab or infliximab, no statistically significant difference in the percentage of patients achieving clinical benefi t was detected between treatment groups or between dose escalators and non–dose escalators, as ascertained by rheumatologists for the medical record subpopulation. Dose escalations did, however, lead to higher TNF blocker costs.

The finding that dose escalation rates are higher among adalimumab and infl iximab users than among etanercept users is consistent with previous research.8,12-21 In previous studies of all 3 agents, the dose escalation pattern was similar to that in the present study, although different treatment periods were investigated and different operational defi nitions of dose escalation were used. The lowest proportion of patients with a dose escalation was found in the etanercept cohort (range, 1%-10%), followed by the adalimumab cohort (range, 8%-34%).14-17,20,21 The greatest proportion was in the infl iximab cohort (range, 7%-35%).14-17,21,26

Despite the higher rates of dose escalation observed among adalimumab and infl iximab users compared with etanercept users in this study, the percentages of patients achieving clinical benefit were similar for the 3 groups. Furthermore, there was no significant difference in the percentage of patients achieving clinical benefit between dose escalators and non–dose escalators within any  treatment cohort. In 2 European studies comparing all 3 TNF-blocking agents in RA patients, higher dose escalation rates were observed among adalimumab and infl iximab users than etanercept users, but changes in disease activity were similar.8,12

In this study, dose escalators had higher therapy costs across all 3 treatment cohorts. This relationship between dose escalation and higher costs is consistent with that observed in previous studies.14,15 When Moots and colleagues12 and Gilbert and colleagues18 compared TNF blocker dose escalators with non–dose escalators, RA-related costs were found to increase significantly only for patients treated with infl iximab or adalimumab. In the current paradigm of comparative effectiveness research, understanding the clinical benefit gained for additional cost is critical to payers.

Study Limitations

Limitations associated with administrative claims data are well recognized. Claims data are collected for the purpose of payment and not research, and are subject to possible errors. Presence of a claim for a filled prescription does not necessarily indicate that the medication was taken as prescribed.

In a responder analysis, a differential dropout rate is a potential bias. The exclusion of patients not on continuous index TNF blocker treatment limits the generalizability of the results. However, results from Khanna and colleagues28 in the same data set show that dropout rates were similar across adalimumab, etanercept, and infliximab cohorts within the first 12 months of initiating biologic treatment. 

The study population consisted of RA patients newly treated with a biologic TNF blocker who were continuously enrolled in a commercial health plan; therefore, results may not be generalizable to other study populations such as Medicare patients or uninsured populations.

Since there is no uniform requirement for physicians’ recording of information in medical records, this process of data collection was associated with missing and incomplete information, and considerable variation was observed in the way information was recorded (ie, different key outcomes were missing from different records). Patients without enough clinical information to evaluate clinical change (eg, fewer than 3 assessments or no key clinical outcome measures) were excluded from medical record analysis. Patients with medical record data for any of the 3 treatments might have had better access to specialty care or more severe RA than the claims population, because rheumatologists might have been more likely to record key clinical outcomes for these patients. Reasons for choosing a particular TNF blocker were not captured, and specific disease or patient characteristics that might have infl uenced response to therapy were unknown.  Reasons for dose escalation were not indicated on most medical records. Information on TNF blockers could not be blinded, since it was within all aspects of the medical record information (eg, infusion or dosing information). We attempted to minimize these limitations in the medical record data by having multiple rheumatologists (at least 4) review the medical records, make the clinical assessment, and agree on final scoring.

Although we controlled for variables in the multivariate analyses that were available in the claims data, certain information such as response rates and disease severity were not indicated on claims and therefore could not be analyzed as confounders.
This study supports earlier research on the comparative effectiveness of tumor necrosis factor (TNF) blockers in rheumatoid arthritis (RA) patients, but is unique in utilizing both clinical outcomes data from US clinical practices and actual drug utilization data from claims. 
  • Increased costs resulting from TNF blocker dose escalation may not always be accompanied by an improvement in clinical benefit. 
  • In the current comparative effectiveness research paradigm, understanding the clinical benefit gained from additional cost is critical to payers.
  • Among RA patients newly treated with adalimumab, etanercept, or infliximab, lower dose escalation rates and costs were seen with etanercept.

Author Affiliations: From Health Economics and Outcomes Research (BC, LB), OptumInsight, Eden Prairie, MN; Sansum Clinic (SDS),  Santa Barbara, CA; Providence Arthritis Center (PB), Portland, OR; Global Health Economics (CW, DJH, SRG), Amgen Inc, Thousand Oaks, CA; K-Force Clinical Research (SC), Tampa, FL.

Funding Source: Research was funded by Immunex Corporation, a wholly owned subsidiary of Amgen Inc, and by Wyeth, which was acquired by Pfizer Inc in October 2009.

Author Disclosures: Mr Chastek and Ms Becker report employment with OptumInsight. Dr Segal reports that he has received lecture fees from Abbott. Dr Bonafede reports receiving consulting and lecture fees from Amgen Inc and Janssen Pharmaceuticals, Inc. Ms Watson reports former employment and stock ownership with Amgen Inc. Mr Chaudhari received payment from Amgen Inc for involvement in the preparation of this manuscript by a contract through K-Force Clinical Research. Drs Harrison and Gandra report employment with Amgen Inc.

Authorship Information: Concept and design (BC, SDS, RPB, CW, DJH, SRG); acquisition of data (BC, SDS, RPB, CW, LB, DJH, SRG); analysis and interpretation of data (BC, SDS, RPB, CW, LB, DJH, SRG); drafting of the manuscript (BC, SDS, RPB, DJH, SRG); critical revision of the manuscript for important intellectual content (BC, RPB, CW, LB, SC, DJH, SRG); statistical analysis (BC, LB, SC); provision of study materials or patients (DJH, SRG); obtaining funding (DJH, SRG); administrative, technical, or logistic support (BC, CW, DJH, SRG); and supervision (BC, CW, DJH, SRG).

Address for correspondence: Benjamin Chastek, MS, Health Economics and Outcomes Research, OptumInsight, 12125 Technology Dr, Eden Prairie, MN 55344. E-mail:
1. Helmick CG, Felson DT, Lawrence RC, et al; National Arthritis Data Workgroup. Estimates of the prevalence of arthritis and other rheumatic conditions in the United States: part I. Arthritis Rheum. 2008;58(1):15-25.

2. Birch JT Jr, Bhattacharya S. Emerging trends in diagnosis and treatment of rheumatoid arthritis. Prim Care. 2010;37(4):779-792.

3. Rindfleisch JA, Muller D. Diagnosis and management of rheumatoid arthritis. Am Fam Physician. 2005;72(6):1037-1047. 4. Maini R, St Clair EW, Breedveld F, 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. ATTRACT Study Group. Lancet. 1999;354(9194):1932-1939.

5. Breedveld FC, Weisman MH, Kavanaugh AF, et al. The PREMIER study: a multicenter, randomized, double-blind clinical trial of combination therapy with adalimumab plus methotrexate versus methotrexate alone or adalimumab alone in patients with early, aggressive rheumatoid arthritis who had not had previous methotrexate treatment. Arthritis Rheum. 2006;54(1):26-37.

6. Weinblatt ME, Keystone EC, Furst DE, et al. Adalimumab, a fully human antitumor necrosis factor alpha monoclonal antibody, for the treatment of rheumatoid arthritis in patients taking concomitant methotrexate: the ARMADA trial [published correction appears in Arthritis Rheum. 2003;48(3):855]. Arthritis Rheum. 2003; 48(1):35-45.

7. van der Heijde D, Klareskog L, Landewé R, et al. Disease remission and sustained halting of radiographic progression with combination etanercept and methotrexate in patients with rheumatoid arthritis. Arthritis Rheum. 2007;56(12): 3928-3939.

8. Blom M, Kievit W, Kuper HH, et al. Frequency and effectiveness of dose increase of adalimumab, etanercept, and infl iximab in daily clinical practice. Arthritis Care Res (Hoboken). 2010;62(9):1335-1341.

9. Abbott Laboratories. Humira (Adalimumab) prescribing information. Published December 20, 2002. Accessed November 8, 2013.

10. Immunex Corporation. Enbrel (Etanercept) prescribing information. Published June 5, 2003. Accessed November 8, 2013.

11. Janssen Biotech. Remicade (Infl iximab) prescribing information. Revised March 2013. Accessed November 8, 2013.

12. Moots RJ, Haraoui B, Matucci-Cerinic M, et al. Differences in biologic doseescalation, non-biologic and steroid intensification among three anti-TNF agents: evidence from clinical practice. Clin Exp Rheumatol. 2011;29(1):26-34.

13. Etemad L, Yu EB, Wanke LA. Dose adjustment over time of etanercept and infliximab in patients with rheumatoid arthritis. Manag Care Interface. 2005;18(4): 21-27.

14. Gu NY, Huang X-Y, Fox KM, Patel VD, Baumgartner SW, Chiou C-F. Claims data analysis of dosing and cost of TNF antagonists. Am J Pharm Benefits. 2010;2(6):351-359.

15. Harrison DJ, Huang X, Globe D. Dosing patterns and costs of tumor necrosis factor inhibit or use for rheumatoid arthritis. Am J Health Syst Pharm. 2010; 67(15):1281-1287.

16. Huang X, Gu NY, Fox KM, Harrison DJ, Globe D. Comparison of methods for measuring dose escalation of the subcutaneous TNF antagonists for rheumatoid arthritis patients treated in routine clinical practice. Curr Med Res Opin. 2010; 26(7):1637-1645.

17. Wu E, Chen L, Birnbaum H, Yang E, Cifaldi M. Retrospective claims data analysis of dosage adjustment patterns of TNF antagonists among patients with rheumatoid arthritis. Curr Med Res Opin. 2008;24(8):2229-2240.

18. Gilbert TD Jr, Smith D, Ollendorf DA. Patterns of use, dosing, and economic impact of biologic agent use in patients with rheumatoid arthritis: a retrospective cohort study. BMC Musculoskelet Disord. 2004;5(1):36.

19. den Broeder AA, Creemers MC, van Gestel AM, van Riel PL. Dose titration using the Disease Activity Score (DAS28) in rheumatoid arthritis patients treated with anti-TNF-alpha. Rheumatology (Oxford). 2002;41(6):638-642.

20. Bullano MF, McNeeley BJ, Yu YF, et al. Comparison of costs associated with the use of etanercept, infliximab, and adalimumab for the treatment of rheumatoid arthritis. Manag Care Interface. 2006;19(9):47-53.

21. Ollendorf DA, Klingman D, Hazard E, Ray S. Differences in annual medication costs and rates of dosage increase between tumor necrosis factor-antagonist therapies for rheumatoid arthritis in a managed care population. Clin Ther. 2009; 1(4):825-835.

22. Harley CR, Frytak JR, Tandon N. Treatment compliance and dosage administration among rheumatoid arthritis patients receiving infl iximab, etanercept, or methotrexate. Am J Manag Care. 2003;9(6)(suppl):S136-S143.

23. Pavelka K, Jarosová K, Suchý D, et al. Increasing the infl iximab dose in rheumatoid arthritis patients: a randomized, double blind study failed to confirm its efficacy. Ann Rheum Dis. 2009;68(8):1285-1289.

24. Alonso-Ruiz A, Pijoan JI, Ansuategui E, Urkaregi A, Calabozo M, Quintana A. Tumor necrosis factor alpha drugs in rheumatoid arthritis: systematic review and metaanalysis of efficacy and safety. BMC Musculoskelet Disord. 2008;9:52.

25. 18. Perdriger A. Infliximab in the treatment of rheumatoid arthritis. Biologics. 2009;3:183-191.

26. Felson DT, Anderson JJ, Boers M, et al; The Committee on Outcome Measures in Rheumatoid Arthritis Clinical Trials. The American College of Rheumatology preliminary core set of disease activity measures for rheumatoid arthritis clinical trials. Arthritis Rheum.1993;36(6):729-740.

27. Boers M, Tugwell P, Felson DT, et al. World Health Organization and International League of Associations for Rheumatology core endpoints for symptom modifying antirheumatic drugs in rheumatoid arthritis clinical trials. J Rheumatol Suppl. 1994;41:86-89.

28. Khanna D, Cyhaniuk A, Bedenbaugh A. Use of TNF inhibitors in the US: utilization patterns and dose escalation from a representative US rheumatoid arthritis population. Poster presented at: EULAR (European League Against Rheumatism) 2011; May 25-28, 2011; London, UK.