Tumor necrosis factor blocker dose escalation among biologic naïve rheumatoid arthritis patients in commercial managed-care plans in the 2 years following therapy initiation

Abstract

Objective:

This study uses real-world US managed-care claims data to estimate dose escalation rates over the first and second years of therapy among biologic naïve rheumatoid arthritis (RA) patients initiating tumor necrosis factor (TNF) blocker therapy with etanercept, adalimumab, or infliximab.

Methods:

Non-elderly adult (age 18–65 years) RA patients initiating etanercept, adalimumab, or infliximab from July 1, 2005 to April 30, 2009, were identified using the MarketScan Commercial Database. National and regional dose-escalation patterns were evaluated 12 and 24 months after initiation. In the single-instance method, dose escalation was defined as having one average weekly dose 115%, 130%, or 150% greater than the initial average weekly dose. By the two-instances method, dose escalation was defined as having two consecutive claims with an average weekly dose 115% or 130% greater than the initial average weekly dose.

Results:

A total of 2747 patients met the inclusion criteria (mean age 50 years [SD = 10]; 74% female). More patients initiated etanercept (44%) than adalimumab (37%) or infliximab (20%). Using the single-instance method, dose escalation at 12 months ranges were 0.8–1.5% for etanercept, 10.8–12.5% for adalimumab, and 16.4–42.5% for infliximab; ranges at 24 months were 0.8–2.1% for etanercept, 14.3–17.5% for adalimumab, and 26.4–57.6% for infliximab. The two-instances method showed a similar relationship among the treatment cohorts at both 12 and 24 months, with lower dose-escalation rates for etanercept (0.8%, 0.8%) than adalimumab (8.7%, 13.3%) or infliximab (22.9%, 37.6%) at the 130% threshold (p < 0.001). Dose-escalation rates for etanercept, adalimumab, and infliximab were consistent across US geographic regions.

Conclusion:

Patients initiating etanercept had lower rates of dose escalation than patients initiating adalimumab or infliximab in the first and second year following therapy initiation, as well as across US geographic regions. These results may not be generalizable to the entire US RA population.

Key words::

• Administrative claims data
• Dose escalation
• Rheumatoid arthritis
• TNF blockers

Introduction

Rheumatoid arthritis (RA) is a chronic inflammatory disorder of unknown etiology that is characterized by pain, inflammation, and progressive joint degeneration1. Biologic therapies that target the proinflammatory cytokine tumor necrosis factor (TNF blockers) have been shown to be effective in the treatment of moderate-to-severe RA.

The three most commonly prescribed TNF blocker therapies are etanercept (Enbrel), infliximab (Remicade), and adalimumab (Humira)2–4; golimumab (Simponi)5 and certolizumab pegol (Cimzia)6 are also FDA-approved for the treatment of RA. Etanercept is a recombinant human soluble TNF receptor protein, and adalimumab and infliximab are anti-TNF monoclonal antibodies. Adalimumab and etanercept are administered as a subcutaneous (SC) injection2,3; infliximab is administered as an intravenous (IV) infusion and is dosed by weight4. For patients with RA, the recommended dose of adalimumab is 40 mg every other week, with the option of increasing the frequency to every week3, and the recommended dose for etanercept is 50 mg per week2. Infliximab is administered as an IV infusion of 3 mg/kg at weeks 0, 2, and 6 and then every 8 weeks thereafter4. The RA dose may be adjusted up to 10 mg/kg or the frequency increased to every 4 weeks4.

Neutralizing antibody formation has been observed with both infliximab7,8 and adalimumab7–11, but not with etanercept12,13. Anti-drug antibodies were reported in 33–43% of RA patients on infliximab therapy7,8 and 12–28% of patients on adalimumab9–11. Presence of anti-infliximab or anti-adalimumab antibodies was associated with a reduced number of clinical responders8,10 or lower mean response9 not occurring with etanercept12,13. Immunogenicity of TNF blockers might play a role in non-response to this treatment10,14. This diminished treatment response may lead to some of the dose escalation seen with these agents in an attempt to regain or maintain clinical response8,15,16; however, the literature reports that higher doses seem to elicit effects similar to lower doses17–20.

Previous studies using US managed-care claims datasets using data from 2002 to 2005 have reported dose-escalation rates of approximately 1–7% for etanercept, 9–21% for adalimumab, and 24–26% for infliximab23–27. Previous studies have shown that patients with dose increases or high starting dose had higher costs compared with patients without high starting dose or dose escalation23,25. In addition to the added costs, increasing the dose of a TNF blocker beyond the minimum recommended label dose increases the risk of side-effects28–30.

While several studies have described TNF blocker dose escalation among RA patients, there is relatively little information regarding dose escalation after the first year of treatment or within specific regions of the US. As studies have indicated, neutralizing antibodies for adalimumab have been detected up to 3 years following therapy initiation. Accounting for long-term dosing patterns is necessary given the chronic nature of RA.

Methods

Study design

A retrospective, observational study was conducted using US administrative claims data to estimate dose-escalation patterns over the first and second years of therapy among biologic naïve RA patients newly initiating TNF blocker therapy with etanercept, adalimumab, or infliximab.

Data source

The study was conducted using administrative claims from the Thomson Reuters MarketScan Commercial Database. The Commercial Database contains the healthcare experience of privately insured individuals covered under a variety of fee-for-service, fully capitated, and partially capitated health plans. The database is constructed from claims and enrollment data provided by over 130 large employer-sponsored health plans from across the US and includes several million enrollees annually, encompassing employees, their spouses, and dependents. There were nearly 40 million enrollees in the Commercial Database in 2009, an increase from approximately 35 million enrollees in 2008. The Commercial Database is de-identified and fully compliant with the Health Insurance Portability and Accountability Act of 1996 (HIPAA). Because this study did not involve the collection, use, or transmittal of individually identifiable data, institutional review board (IRB) review or approval was not required.

Study inclusion criteria

Non-elderly adults with RA, ages 18–65 years, newly initiating TNF blocker therapy with infliximab, etanercept, or adalimumab during the time period from July 1, 2005, through April 30, 2009, were selected for the study. RA patients were identified by the presence of at least one inpatient or two outpatient medical claims containing the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) code 714.0x. The index date was identified as the date of the first TNF blocker prescription or administration on or after January 1, 2005; the index medication was the TNF blocker prescribed or administered on the index date. Data through April 30, 2010, was used to observe dose-escalation patterns. Patients were required to have 6 months of continuous medical and pharmacy enrollment (i.e., the pre-index period) that was clean of any TNF blocker use, including the index medication.

Patients were required to be continuously enrolled and persistent on their index therapy for 12 months following therapy initiation. Patients were considered to be non-persistent if they switched to a non-index TNF blocker or if they had a gap in their index therapy of greater than 60 days. Dose escalation was evaluated at 24 months following initiation for patients who were continuously enrolled and persistent on index therapy for the entire 2 years. By definition, patients persistent on index therapy for 24 months are a subset of patients who were persistent on their index therapy for 12 months. Patients who were persistent for greater than 12 months but less than 24 months were included only in the 12-month cohort.

Study exclusion criteria

Patients were excluded from the study if they were under 18 or over 65 years of age on the index date. Patients were excluded if they used any of the following medications during the pre- or post-index periods: anakinra (Kineret), rituximab (Rituxan), abatacept (Orencia), tocilizumab (RoActemra, Actemra), certolizumab pegol (Cimzia), or golimumab (Simponi); there was insufficient sample to include these therapies as stand-alone treatment cohorts during the study period. Patients were also excluded if they had the following diagnoses during the pre- or post-index periods: non-Hodgkin’s lymphoma (ICD-9-CM: 200.xx or 202.xx), chronic lymphocytic leukemia (ICD-9-CM: 204.10), psoriasis (ICD-9-CM: 696.1), psoriatic arthritis (ICD-9-CM: 696.0), juvenile idiopathic arthritis (ICD-9-CM: 714.3), Crohn’s disease (ICD-9-CM: 555.x), ulcerative colitis (ICD-9-CM: 556.x), ankylosing spondylitis (ICD-9-CM: 720.0), cancer (ICD-9-CM: 140.xx–209.xx or V10), or HIV (ICD-9-CM: 042). Patients with poor quality claims data (i.e., missing or zero values for quantity or days supply) or outlier doses on any claims (<50% or >200% of labeled dose) were excluded. Lastly, this study was limited to patients receiving greater than or equal to the recommended label dose as their initial dose; patients initiating therapy with suboptimal doses were excluded to ensure that titrating up to the recommended dose was not misclassified as dose escalation.

Outcome variables

Due to the claims database structure, dose was calculated differently for subcutaneously injected and infused products. For etanercept and adalimumab, the initial dose was the dose of the index prescription. The average weekly dose per patient was calculated using the total quantity on the claim divided by the total days’ supply. Since infliximab is infused, it is billed as a medical service and not a pharmacy claim, requiring imputation to determine the dose per claim. The infliximab dose was calculated using the paid amount on the infliximab claim divided by the wholesale acquisition cost (WAC) for that calendar year. Infliximab administrations are scheduled at weeks 0, 2, 6, and then every 8 weeks thereafter; therefore, the initial dose was based on the fourth infusion, which is the first infusion following the loading period.

Dose escalation was evaluated using two separate methods, both of which are based on the average weekly dose for each claim. Under the first method, the single-instance method, dose escalation was defined as having an average weekly dose on a claim that is greater than 115% of the initial claim’s average weekly dose; the percentage of patients with an average weekly dose greater than 130% or 150% of the initial claim was also reported. Under the second method, the two-instances method, dose escalation was defined as having two consecutive claims where the average weekly dose on each claim is greater than 115% of the initial claim’s average weekly dose; a separate threshold of 130% was also used for the two-instances method.

Using both methods, dose escalation was reported separately by geographic region of the US in the 12 months after therapy initiation; national estimates of dose escalation using both methods were reported in the 12 and 24 months after therapy initiation. Geographic regions included the Northeast, South, North Central (including Midwest) and West.

Descriptive analysis

Descriptive statistics were used to summarize the characteristics of patients in the three TNF blocker cohorts. Chi-squared tests were used to test for differences in nominal/categorical variables; t-tests were used to test for differences in interval/continuous variables. Etanercept was the reference group, with an a priori significance level of 0.05. Analysis was performed using SAS version 9.2 (SAS Institute, Cary, NC, USA)31.

Results

Baseline demographics and clinical characteristics

A total of 2747 RA patients newly initiating etanercept, adalimumab, or infliximab between July 1, 2005, and April 30, 2009, met all of the study selection criteria. Figure 1 contains a patient attrition diagram.

Figure 1.  Patient attrition.

Of those meeting the study criteria, 1207 (44%) initiated treatment with etanercept, 1004 (37%) with adalimumab, and 536 (20%) with infliximab. A total of 1125 patients were persistent on their index therapy 24 months following therapy initiation; 447 (42%) of these patients initiated etanercept, 399 (35%) initiated adalimumab, and 250 (22%) initiated infliximab. Patient characteristics were similar across the three treatment groups. Among patients with at least 12 months of follow-up, the mean age was 50 (SD = 10 years), the majority of patients were female (73–75%), and the majority belonged to a non-capitated health plan (79–82%). The highest proportion of patients lived in the South (44.6%), followed by the North Central (28.2%), West (16.5%), and Northeast (10.2%). Table 1 contains patient demographic and clinical characteristics stratified by index medication, region, and length of index medication persistence (12 or 24 months). Baseline Deyo Charlson Comorbidity Index (DCI) scores were similar across treatments, ranging from 1.07 (SD = 0.29) among adalimumab patients to 1.13 (SD = 0.50) among infliximab patients. Methotrexate (71–75%) and glucocorticoids (57–61%) were the most prevalent concomitant medications used during the pre-index period.

Table 1.  Demographic and clinical characteristics of patients persistent with etanercept. Adalimumab, or infliximab far 12 or 24 months.

Initial dose

The average initial weekly dose was 48.7 mg (SD = 2.1 mg) for etanercept, 21.0 mg (SD = 4.8 mg) for adalimumab, and 52.6 mg (SD = 24.3 mg) for infliximab. The majority of patients initiated TNF blocker treatment at the recommended label dose. Nearly all etanercept and adalimumab patients (99% and 93%, respectively) and 72% of infliximab patients initiated therapy within 10% of the recommended label dose; remaining patients initiated therapy with a dose at least 10% higher than the recommended label dose.

Dose escalation

Table 2 contains the dose-escalation results for both the single- and two-instances methods. Under the single-instance method, patients in the adalimumab (12.5%) and infliximab (42.5%) cohorts were both more likely than etanercept (1.5%) patients to have an average weekly dose that was greater than 115% of the initial dose (p < 0.001 for both comparisons). Under the single-instance method, patients in the adalimumab (12.0%) and infliximab (30.2%) cohorts were both more likely than etanercept (1.2%) patients to have an average weekly dose that was greater than 130% of the initial dose (p < 0.001 for both comparisons); 0.8% of etanercept patients, 10.8% of adalimumab patients, and 16.4% of infliximab patients had dose escalation in the 12 months following therapy initiation using the 150% threshold (p < 0.001 for adalimumab versus etanercept and infliximab versus etanercept, separately). Using the two-instances method, 0.8% of etanercept patients had dose escalation using the 130% threshold in the 12 months following therapy initiation, compared with 8.7% of adalimumab patients and 22.9% of infliximab patients (p < 0.001 for both comparisons). For patients persistent on their index therapy for 24 months, 1.5% of etanercept patients had dose escalation using the 130% threshold for the single-instance method, compared with 16.5% of adalimumab patients and 41.2% of infliximab patients (p < 0.001 for both comparisons); 0.8% of etanercept patients had dose escalation in the 24 months following therapy initiation at the 150% threshold compared with 14.3% of adalimumab patients and 26.4% of infliximab patients (p < 0.001 for both comparisons). Using the two-instances method, 1.0% of etanercept patients had dose escalation in the 24 months following therapy initiation at the 130% threshold, compared with 13.8% of adalimumab patients and 51.2% of infliximab patients (p < 0.001 for both comparisons).

Table 2.  Dose escalation in the first 24 months of etanercept, adalimumab, or infliximab use.

	12-Months						24-Months
	Etanercept		Adalimumab		Infliximab		Etanercept		Adalimumab		Infliximab
	N = N/Msan	1.207%/SD	N = N/Mean	1.004%/SD	N = N/Mean	536%/SD	N = N/Mean	477%/SD	N = N/Mean	399%/SD	N = N/Mean	250%/SD
Index weekly dose (Mean, SD)	48.7	2.1	21.0	4.8	52.6	24.3	48.8	2.6	21.1	4.7	52.4	22.2
Percentage with dose increase from index (N, %)
115%	13	1.5%	126	12.5%	223	42.5%	10	2.1%	70	17.5%	144	57.6%
130%	14	1.2%	120	12.0%	162	30.2%	7	1.5%	66	16.5%	103	41.2%
150%	10	0.8%	103	10.8%	33	16.4%	4	0.8%	57	14.3%	66	26.4%
Two instances method (N, %)
115%	11	0.9%	91	9.1%	134	34.3%	5	1.0%	55	13.8%	123	51.2%
130%	10	0.8%	87	8.7%	123	22.9%	4	0.8%	53	13.3%	94	37.6%

Note: All differences between etanercept and adalimumab and between etanercept and infliximab are significant at p < 0.001.

Regional analysis of dose escalation

Rates of dose escalation by geographic region were similar to the national rates. Using the single-instance method and 150% threshold, dose-escalation rates for etanercept ranged from 2.3% in the West to 3.9% in the Northeast in the 12 months following therapy initiation. Using the same criteria, dose-escalation rates ranged from 8.5% in the West to 13.1% in the Northeast for adalimumab and from 26.5% in the Northeast to 39.5% in the West for infliximab. Figure 2 depicts regional dose-escalation rates.

Figure 2.  Dose escalation in the 12 months following therapy initiation, using the single-instance method at 130% of the initial dose.

Discussion

This study is a comprehensive assessment of dose escalation among patients initiating TNF blockers, providing recent estimates of dose escalation among biologic naïve RA patients newly initiating TNF blocker treatment with employer-paid commercial healthcare insurance in the 12 and 24 months following treatment initiation. This study used two methods and multiple thresholds for defining dose escalation. Using the most stringent definition (two instances, 130% threshold) and least stringent definition (single instance, 115% threshold), dose-escalation rates ranges were 0.8–1.5% for etanercept, 8.7–12.5% for adalimumab, and 22.9–42.5% for infliximab in the year following therapy initiation. This is similar to a previous study evaluating dose escalation among TNF blocker naïve RA patients24, which reported that dose escalation was lowest among patients treated with etanercept and highest among those treated with infliximab. Harrison and colleagues used commercial healthcare claims data from January 2004 to December 2005 to evaluate dose escalation among biologically naïve RA patients initiating TNF blocker therapy and reported that 2% of etanercept patients had a dose increase during the 12-month observation period compared with 14% of adalimumab patients and 60% of infliximab patients24. Harrison defined dose escalation as having a subsequent dose that is 10% greater than the index dose, which may explain why Harrison’s reported rates are slightly higher than this current analysis. Other published studies have shown higher dose-escalation rates in the 12 months following therapy initiation. Huang and colleagues used multiple methods for evaluating dose escalation and reported dose-escalation rates in the ranges 5.5–10.3% for etanercept patients and 14.8–33.6% for adalimumab patients using the same, albeit older, managed-care data source as this current analysis26. Using a large national managed-care database from April 2007 through March 2009, Khanna and colleagues found that 3% of etanercept patients increased their weekly dose from 50 mg to 100 mg, 13% of adalimumab patients increased their biweekly dose from 40 to 80 mg, and 39% of infliximab patients either added another vial or reduced the time between infusion administrations to less than 6 weeks32. Some of the variation in the dose-escalation estimates in previous research may be due to the method of calculating dose escalation, differences in the patient population, or differences in trends over time. However, despite these differences in treatment period and operational definitions of dose escalation, there is a consistent trend of higher rates of dose escalation among adalimumab and infliximab patients compared with etanercept patients.

Dose escalation is an important outcome due to its relationship with costs and the similar percentage of patients achieving clinical improvements with increased dose. Previous studies have reported a relationship between dose escalation and higher costs23,24, while other studies have shown that treatment groups with higher rates of dose escalation have increased therapy costs, whereas low population rates of dose escalation are associated with stable costs25^,33. Blom17 and Moots18 both reported higher dose-escalation rates among adalimumab and infliximab users compared with etanercept users; however, changes in disease activity were similar across treatment groups. Likewise, Schabert and colleagues found similar disability outcomes among US patients in an RA registry treated with etanercept, adalimumab, or infliximab, even though self-reported dose escalations of the index medication were observed more often among infliximab and adalimumab patients than among etanercept patients21. In a study using rheumatologist chart assessment data, patients on etanercept were less likely to have dose escalation compared with patients on adalimumab or infliximab; however, there were no statistically significant differences in percentage achieving clinical improvement between treatment groups or between dose escalators and non-escalators22.

While previous studies have reported dose escalation in the first 12 months following therapy initiation, there is relatively little data regarding dose escalation over longer time periods. Given the chronic nature of TNF blocker use and the fact that secondary resistance to TNF blocking agents can develop, a longer time period is useful for better understanding dose-escalation rates in the real world. This study found that dose escalation with etanercept was relatively stable from the first to second year, increasing by 0.3% (from 1.2 to 1.5%) using the single-instance method and 130% threshold. By comparison, dose escalation increased from 12.0 to 16.5% for adalimumab patients and from 30.2 to 41.2% for infliximab patients using the same measure.

This study also provides regional estimates of dose escalation, which can be used by managed-care organizations to compare their dose-escalation rates to regional and national benchmarks and to inform formulary decisions. Results of the regional analysis were directionally consistent with the national dose-escalation rates. Dose-escalation rates among etanercept patients were low compared with adalimumab and infliximab and remained stable across all four geographic regions. Dose-escalation rates for adalimumab and infliximab showed more regional variation, but were directionally consistent with national dose-escalation rates.

Limitations

This study was subject to several limitations. Newer biologic RA therapies (such as certolizumab and golimumab) were not included in this analysis due to insufficient sample size within the study time period. While requiring patients to be continuously enrolled and persistent on therapy for 12 (or 24) months created a homogenous patient population and allowed for equitable comparisons of dose escalation across treatments, they also limit the generalizability of the study findings. For example, the results of this analysis may not apply to patients with shorter treatment durations or characteristics that may lead to unstable insurance coverage. Similarly, the data source includes only patients with commercial health coverage; therefore, the results may not be generalizable to the entire US RA population. The data source is also limited to information found on administrative healthcare claims; disease-specific clinical information (e.g., disease remission status, disease-related quality of life, etc.) could not be included in this analysis.

Conclusion

RA patients newly initiating etanercept had significantly lower rates of dose escalation compared with patients initiating adalimumab or infliximab (p < 0.001 for all comparisons). This finding was consistent across dose-escalation measures in both the first and second year following therapy initiation, as well as across geographic regions of the US. These findings should be taken into consideration while making formulary decisions regarding TNF blocker therapy, given the lack of evidence of greater effectiveness of dose escalating agents and the fact that increased dose is associated with higher costs and increased risk of adverse events23,25,28–30.

Transparency

Declaration of funding

This 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.

Declaration of financial /other relationships

S.G. is an employee of Amgen, Inc. and has received Amgen stock/stock options; M.B. and K.W. are employees of Thomson Reuters, which received a research contract to conduct this analysis. K.F. received research funds from Amgen, Inc. as a consultant.

Acknowledgments

The peer reviewers on this manuscript have disclosed any relevant financial relationships.