Claims Data Analysis of Dosing and Cost of TNF Antagonist

, , , , ,
The American Journal of Pharmacy Benefits, Specialty Pharmacy, Volume 2, Issue 6

Higher drug costs may be attributable to either high initial doses or a dose increase.

Etanercept (Enbrel), adalimumab (Humira), and infliximab (Remicade) are the commonly prescribed tumor necrosis factor-alpha (TNF-alpha) antagonists for rheumatoid arthritis (RA) in routine clinical practice. These agents are indicated for reducing the signs and symptoms of RA, inhibiting the progression of structural damage, and improving physical function in patients with moderate to severe RA.1-3 These 3 products have additional indications other than RA, including ankylosing spondylitis, plaque psoriasis, polyarticular-course juvenile idiopathic arthritis, and psoriatic arthritis for etanercept; polyarticular-course juvenile idiopathic arthritis, ankylosing spondylitis, Crohn’s disease, psoriatic arthritis, and plaque psoriasis for adalimumab; and ankylosing spondylitis, psoriatic arthritis, plaque psoriasis, Crohn’s disease, and ulcerative colitis for infliximab.

The etanercept dose recommended by the US Food and Drug Administration (FDA) for RA is 50 mg per week administered as a single 50-mg subcutaneous (SC) injection or two 25-mg SC injections.1 The FDA-recommended dose of adalimumab for RA is 40 mg every other week as 1 subcutaneous injection, with the option to increase the dosing frequency to 40 mg every week in patients who are not concurrently taking methotrexate.2 The FDA-recommended dose of infliximab for RA is 3 mg/kg, administered as an intravenous infusion followed by additional infusions at 2 and 6 weeks after the first infusion and then every 8 weeks thereafter with the option to adjust the dose upward to 10 mg/kg as often as every 4 weeks.3

Clinical trials have established the safety and efficacy of the anti-TNF-alpha agents in the treatment of RA.4-6 However, not all patients respond adequately to the recommended starting dose. It was reported that the most common reason for increasing the dose of infliximab was insufficient clinical response.7 St. Clair and colleagues noted breakthrough joint pain and swelling several weeks prior to the next infliximab infusion in up to 25% of patients.8

A potential mechanism for reduced efficacy of a biologic is the development of antibodies to the drug.9 In clinical studies, approximately 6% of etanercept-treated subjects with RA, psoriatic arthritis, ankylosing spondylitis, or plaque psoriasis developed antibodies to etanercept.1 However, no neutralizing antibodies (ie, antibodies that neutralize the function of the drug) were detected and no correlation between antibody production and clinical response or safety was noted.1 In contrast, the incidence of anti-infliximab antibodies was approximately 10% in subjects receiving infliximab for up to 2 years.3 Patients who were antibody positive were more likely to have higher rates of clearance and reduced efficacy, and to experience an infusion reaction.3 According to the adalimumab product label, approximately 5% of adult RA patients developed low-titer antibodies to adalimumab at least once during treatment, which were neutralizing in vitro.2 In a 2007 study, Bartelds et al found the presence of anti-adalimumab antibodies in RA subjects receiving adalimumab monotherapy was correlated with reduced serum concentration and reduced clinical response.10

Higher-than-recommended doses may result in higher costs for payers and potentially for patients through higher co-payments or coinsurance. Ollendorf et al examined dosing patterns and costs among RA patients newly treated with infliximab and found that annualized RA-related costs were higher by more than 50% among patients with upward dose adjustment.11 Hence, agents that achieve and maintain clinical response at the recommended starting dose should be less costly to payers and patients.

A review article by Ariza-Ariza et al revealed that a large proportion of infliximab patients underwent dose escalation, whereas the mean dose increase of the etanercept was not statistically significant.12 Dose escalation for adalimumab patients was not reported because no studies on adalimumab met the inclusion criteria.12 Without dosing information on infliximab, the objective of this study was to investigate the rate of dose increases for etanercept and adalimumab and direct costs for all 3 biologics.


A cohort study was conducted to assess dosing patterns and costs associated with TNF-alpha antagonists among RA patients with newly initiated TNF-alpha antagonist treatment. The Medstat MarketScan Commercial Claims and Encounters and Medicare Supplemental and Coordination of Benefits (COB) databases from January 1, 2002, through December 31, 2004, were utilized for this study. The Medstat databases capture person-specific clinical utilization, expenditures, and enrollment across inpatient, outpatient, prescription drug, and carve-out services from approximately 45 large employers, health plans, and government organizations. Specifically, the Commercial Claims and Encounters database provides data for active employees, early retirees, and their dependents who are insured by employer-sponsored plans. The Medicare Supplemental and COB database provides information about Medicareeligible retirees with employer-sponsored Medicare supplemental plans. Historically, more than 500 million claim records are available in the MarketScan databases.

July 1, 2002, through December 31, 2003, was established as the study intake period for identifying the initiation of anti-TNF-alpha therapy. The 6-month period preceding the start of anti-TNF-alpha therapy was used as a baseline to ensure that the patient was naïve to RA biologics. The 12-month period following initiation of anti-TNF-alpha therapy was used to assess dosing patterns and total cost of care.

Patients were eligible for the study if they (1) had at least 1 diagnosis of RA (International Classificationof Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] codes 714.0, 714.1, 714.2 or 714.81); (2) had at least 2 claims for the same biologic; (3) had continuous enrollment for 6 months prior to the first biologic claim (baseline period); and (4) had 12 months continuous enrollment following the first biologic claim (follow-up period). Patients were excluded from the study if they met any of the following criteria: (1) had any claim for biologics of interest in the baseline period; (2) had a diagnosis of Crohn’s disease (ICD-9-CM codes 555.0, 555.1, 555.2, 555.9, 565.1, 569.81), psoriasis or psoriatic arthritis (ICD-9-CM codes 696.0, 696.1, 696.8), or ankylosing spondylitis (ICD-9-CM code 720.0); or (3) had claims for 2 or more different biologics during the study period.

Dosing Pattern Assessment

For adalimumab and etanercept, the weekly dose was calculated from the days supply field (converted into weeks) and dose (converted into milligrams) in the pharmacy claims. For infliximab, the unit of service was not available in the medical claims to verify the number ofvials (per Healthcare Common Procedure Coding SystemJ codes); hence, dosing pattern and dose increase couldnot be assessed for infliximab patients.

To quantify dose increases over time, an exposure-adjusted approach was used. Specifically, dose increase was estimated at each week based on the number of patients treated with a given biologic (denominator) and the number of patients with a dose increase greater than the weekly dose of the index prescription (numerator). To capture the overall trend of the dosing patterns, cumulative rates of dose increase were reported. Because

the starting dose affects the rate of the dose increase,

analyses were conducted separately for patients with a

starting dose less than or equal to the recommended dose

or greater than the recommended dose.

Healthcare Costs

In this study, costs were assessed from payer and patient perspectives by utilizing paid claims. For Medicare patients, both the Medicare-paid amount and the employer-paid supplemental insurance amounts were included. Net payment for each of the services (pharmacy, inpatient, and outpatient) was compared across the anti-TNF-alpha agents to reflect a health plan’s perspective. Similarly, patients’ out-of-pocket costs for each of the services were examined separately to determine the patient cost burden. Pharmacy costs were defined as total pharmacy costs for all drugs prescribed to the RA patients, including the biologics. Inpatient and outpatient costs included net payment for hospitalizations, emergency department visits, physician office visits, procedures, and laboratory tests. Patient’s out-of-pocket expense consists of coinsurance, co-payments, and deductibles. Costs were aggregated over 12 months of follow-up.

Statistical Analysis

Descriptive statistics for eligible patients by treatment group were provided. The Wilcoxon rank-sum test was used for continuous measures, and the c2 test was used for discrete measures. Statistical comparisons were conducted using etanercept as the reference cohort. For mean values, nonparametric analysis of variance was used to detect statistically significant differences in costs between the anti-TNF-alpha agents. For median values, the Kruskal-Wallis test was used to detect statistically significant cost differences among the anti-TNF-a agents. All statistical analyses were performed using SAS version 9.1.3 (SAS Institute Inc, Cary, NC).


A total of 2320 patients were eligible for the study (

Figure 1

). Etanercept patients (n = 1369, mean age 50 years [SD 10.9 years]) were slightly younger than adalimumab patients (n = 461, mean age 52 years [SD ± 9.4 years]) and infliximab patients (n = 490, mean age 52 years [SD 9.6 years]). With regard to sex, comorbidity,13 and insurance coverage, etanercept patients were similar to adalimumab patients (P >.05). On the other hand, etanercept patients differed from infliximab patients in terms of geographic location and insurance coverage (P<.001) (

Table 1


Dosing Pattern and Associated Cost

The mean weekly dose was 50.5 mg for etanercept and 22.3 mg for adalimumab. With regard to starting dose, 11.9% (n = 55) of adalimumab patients began at >20 mg per week, compared with 2.8% (n = 38) of etanercept patients whose starting dose was >50 mg per week (P <.001) (Table 1).

Patient attrition over the 52 weeks was similar for etanercept and adalimumab patients, so the patterns of dose increase could be compared. Average treatment duration was 34.7 weeks for etanercept and 33.0 weeks for adalimumab (P = .03), reflecting the fact that patients who initiated treatment with etanercept continued therapy slightly longer than those who initiated treatment with adalimumab (Table 1).

Results of the exposure-adjusted method showed that etanercept and adalimumab patients had similar rates of dose increase during early weeks of treatment (weeks 1-12) and a divergent pattern of dose increase throughout the rest of the follow-up (

Figure 2

). This overall trend of dosing patterns for adalimumab and etanercept was demonstrated in their upward sloping over the 12-month follow-up. A steeper slope was shown for patients treated with adalimumab than for those treated with etanercept. By about week 21, the proportion of adalimumab patients with a dose increase was almost twice that of etanercept patients. By the end of the follow-up, about 8% of adalimumab patients who started at the recommended dose had a dose increase, as opposed to 3.4% of etanercept patients. Further, patients on adalimumab who started on a high dose (11.9%) continued on the high dose throughout the treatment period. A total of up to 20.7% of adalimumab patients were at a dose of >20 mg weekly throughout the treatment period compared with 7.1% of etanercept patients (

Figure 3


Although dose was not available for infliximab patients, the costs of care were available, and the mean total annual costs for the three biologics were estimated (

Table 2

). Average total direct costs (including all pharmacy, inpatient, and outpatient care) were $17,236 for infliximab patients, $16,086 for adalimumab patients, and $14,490 for etanercept patients annually. Hence, the average cost difference between adalimumab and etanercept was $1595 (P <.01). The average total costdifference between infliximab and etanercept was $2746(P <.001). The annual average cost of biologics per patient (included in the total pharmacy cost as shown inTable 2), was $13,630 for infliximab patients, $12,914for adalimumab patients, and $11,882 for etanerceptpatients. The average cost difference for biologics betweenadalimumab and etanercept was $1032 (P <.01),and the average cost difference between infliximab andetanercept was $1748 (P <.001). Distributions of costswere skewed as expected; hence, median costs also areprovided in Table 2.

Figure 4

illustrates the average amount paid by patients during the follow-up period ($1421, $874, and $755 for infliximab, adalimumab, and etanercept patients, respectively). Patients receiving infliximab paid significantly more (P <.001) than adalimumab and etanercept patients (Figure 4).

The impact on cost of care for patients who had a dose increase compared with patients who did not have a dose increase or who had a high starting dose is shown in

Figure 5

. When comparing the cost for patients who had a high starting dose or a dose increase with the cost for patients who did not have a high starting dose or a dose increase, it was found that the cost difference was significantly larger for the adalimumab group than the etanercept group ($4154 vs $683, P <.001) (Figure 5).


Analyzing dose patterns using administrative claims data is important because it provides information regarding the use patterns of a drug in the routine clinical setting as well as the associated treatment cost.12 Although several methods have been used in other studies to measure dose increase, including comparing the last prescription with the first prescription and 2 instances of a higher-dose prescription than the index prescription,14-17 the exposure-adjusted approach was utilized because it could capture the dynamic nature of dose increase. Because the exposure-adjusted method provided weekly incidences of dose increase as well as the pattern of dose increase over time, divergence or convergence could be observed. This is important clinically because it indicates when prescribers made the decision to increase the dose. From a cost perspective, the timing of dose increases indicates how soon costs will escalate after therapy initiation. Further, biologic therapies are very costly. Doan et al reported in their review of the economic value of biologic disease-modifying antirheumatic drugs (DMARDs), “Biologic therapies are more costly compared with traditional DMARDs but produce more QALYs.”18 It is important economically to evaluate the cost impact of biologic DMARDs in relation to treatment options for management of RA patients either resistant to or intolerant of traditional DMARDs.

Because the dose of the index prescription is frequently used as the basis to calculate the dose increase, it is important to document the starting dose. In this study, 11.9% of adalimumab patients and 2.8% of etanercept patients started on a dose greater than the recommended starting dose. As would be expected, dose increases were less frequent in those patients. It should be mentioned that adalimumab >40 mg per week is an FDA-approved dose, whereas etanercept >50 mg per week is not. Nonetheless, this study reflects the higher costs associated with the higher percentage of patients who require the higher dose of anti-TNF-alpha agents.


Administrative claims data are used extensively today to examine real-world treatment patterns, outcomes, and costs of care. However, this study has the following limitations:

• The presence of a claim for a filled prescription does not indicate that the medication was actually consumed by the patient, nor does it indicate that the medication was taken as prescribed.

• No assessment can be made of the reasons for dose increases.

• To enhance robustness of the study findings, patients who switched to other biologics during the 12-month follow-up were excluded. Therefore, the study is similar to a responder analysis, but clinical response information is not available.

• Because patients were not randomly assigned to different biologic treatments, the findings may be confounded by differences between treatment groups in baseline disease severity (selection bias) or other unmeasured factors.

• Because adalimumab has 2 approved dose regimens for RA, patients receiving adalimumab are more likely to have a dose increase than patients receiving etanercept. According to the adalimumab product label, the recommended dose is 40 mg every other week. However, for some patients not taking concomitant methotrexate, the dosing frequency may be increased to 40 mg every week.

• Most managed care plans are likely to have prior authorization criteria that may limit the extent of dose increase. Plans without these criteria may have more extensive dose increases. Hence, our estimates of dose increase may be conservative. On the other hand, actual use of these drugs in plans with prior authorization criteria is likely to closely follow the product labeling.

• The data set we used was from 2002-2004. Information we extracted was representative of the plans included in the analysis, and may not be generalizable to the rest of the US population.

• Cost adjustment was not made despite the use of claims from both 2003 and 2004.

• Patient-paid cost may reflect benefit design and formulary status, in addition to dosing.

• Caution should be exercised in generalizing our findings beyond a 1-year time frame because the therapeutic regimen may be altered beyond this period.

• As previously stated, due to lack of information, we were unable to assess dose increase for infliximab patients.


Among RA patients newly treated with biologics, 11.9% of adalimumab and 2.8% of etanercept patients were started on a dose higher than the recommended dose. Although adalimumab and etanercept patients had similar rates of dose increase in the first 12 weeks of treatment, adalimumab patients had a significantly higher rate of dose increase than patients on etanercept during the rest of the 12-month follow-up period. In this population, it appears that patients with a high starting dose or a dose increase had statistically significant higher total costs compared with patients who did not have a high starting dose or a dose increase.

Based on the dosing patterns illustrated in this study, there appears to be an association between the use of adalimumab and higher doses, which also may be related to the higher pharmacy cost for adalimumab compared with etanercept. At the same time, given the study limitations, these results should be interpreted with caution in managed care decision making.