Dose optimization of infliximab in patients with rheumatoid arthritis

Authors

  • Rieke Alten,

    Corresponding author
    1. Department of Internal Medicine II, Rheumatology, Clinical Immunology, Schlosspark-Klinik Teaching Hospital, Charité University Medicine Berlin, Berlin, Germany
    • Correspondence: Rieke Alten, Head of Department of Internal Medicine II, Rheumatology, Clinical Immunology, Osteology Schlosspark-Klinik, Charité–University Medicine Berlin, Heubnerweg 2, 14059 Berlin, Germany.

      Email: rieke.alten@schlosspark-klinik.de

    Search for more papers by this author
  • Filip van den Bosch

    1. Department of Rheumatology, Ghent University Hospital, Ghent, Belgium
    Search for more papers by this author

Abstract

Aim

It is well established that tumor necrosis factor (TNF) inhibitors help control disease activity, limit radiographic progression and preserve function in patients with rheumatoid arthritis. However, not all patients respond adequately to initial anti-TNF treatment and some patients lose response over time. Possible treatment modifications include optimizing concomitant disease-modifying antirheumatic drugs, switching to another anti-TNF biologic or another class of agent, or optimizing the dose of the anti-TNF agent. Here we review data on dose optimization of infliximab, with emphasis on dose changes to address inadequate response, nonresponse and loss of response.

Method

The authors conducted a literature review to identify studies that evaluated the effect of dose optimization on clinical response in infliximab-treated patients with rheumatoid arthritis.

Results

Few well-controlled studies of dose optimization of infliximab have been completed for patients with rheumatoid arthritis, and the evidence supporting efficacy and safety after dose adjustment can be difficult to interpret. Studies of dose optimization in infliximab-treated patients who fail to show initial response, have an inadequate response, or lose response over time are not entirely consistent, but tend to show a pattern of improvement after a dose increase.

Conclusion

Dose optimization involves a balance of risks and benefits, and future research should seek to clarify which patients are most likely to benefit from dose optimization without undue increase in risk.

Introduction

In patients with rheumatoid arthritis (RA) who have failed other synthetic disease-modifying antirheumatic drug (DMARD) treatments, tumor necrosis factor (TNF) inhibitors are an important treatment option. TNF inhibitors, usually combined with methotrexate (MTX) and/or other DMARDs, have been shown to improve radiographic, clinical and functional outcomes in patients with RA.[1-3] However, not all patients respond optimally to initial treatment, and in these cases several treatment adjustments might be considered. Optimizing concomitant DMARDs, such as MTX, is the first step in modifying treatment. Adjustment of biologic therapy may then be considered, including switching anti-TNF agents; moving to a different medication class altogether (e.g., abatacept, rituximab); or continuing with the same agent and optimizing the dose. Current international guidelines for RA advise treatment to target, and the latest recommendation is that drug therapy should be adjusted at least every 3 months until the treatment target is reached.[4]

Dose optimization, or modification of the drug regimen to achieve maximum effectiveness while minimizing risk, may be desirable in several circumstances during treatment with TNF antagonists in patients with RA. Dose escalation may be appropriate when a patient experiences non-response, inadequate response, disease flare or when a patient who initially responded loses response over time.[5-7] Dose tapering may also be advisable to achieve the minimal effective dose after low disease activity has been reached[8] or when safety issues arise. As part of a tight-control strategy, doses may be optimized based on frequent assessment of therapeutic outcomes to achieve treatment goals.[9, 10]

There is no clear and comprehensive guidance regarding dose optimization for TNF antagonists, in part due to lack of rigorous data available to guide best practice recommendations.[6] Product labeling of TNF inhibitors includes a recommended dose range for some, but not all, of the TNF inhibitors (Table 1).[11-20] The US product labeling of infliximab, for example, states a maintenance dose of 3 mg/kg, and dose increases for non-response up to 10 mg/kg every 8 weeks or frequency increases up to every 4 weeks. However, guidelines vary across countries. European guidelines recommend a maximum infliximab dose of 7.5 mg/kg every 8 weeks or 3 mg/kg every 4 weeks.[12]

Table 1. Dosing of TNF inhibitors in rheumatoid arthritis according to US and EU Prescribing Information
TNF InhibitorUS Prescribing InformationEU Summary of Product Characteristics
Maintenance doseFrequencyPossible dose adjustmentMaintenance doseFrequencyPossible dose adjustment
  1. EU, European Union; IV, intravenous; N/A, not available; RA, rheumatoid arthritis; SC, subcutaneous; TNF, tumor necrosis factor; US, United States.

  2. a

    For RA, must be given concomitantly with methotrexate.

Infliximab,a IV[11, 12]3 mg/kgEvery 8 weeksFor loss of response, increase dose up to 10 mg/kg or treat as often as every 4 weeks3 mg/kgEvery 8 weeksFor inadequate or loss of response after 12 weeks, dose may be increased by 1.5 mg/kg increments up to 7.5 mg/kg or 3 mg/kg may be administered as often as every 4 weeks
Adalimumab, SC[13, 14]40 mgEvery 2 weeksSome RA patients not receiving methotrexate may benefit from increasing frequency to 40 mg every week40 mgEvery other weekSome patients receiving monotherapy may benefit from increasing frequency to 40 mg every week
Etanercept, SC[15, 16]50 mgOnce per weekN/A25 mgTwice weekly50 mg once weekly may be administered as an alternative dosing regimen
Golimumab, SC[17, 18]50 mgOnce per monthN/A50 mgOnce per month

In patients weighing more than 100 kg with inadequate response after 3 or 4 doses, increase dose to 100 mg once a month

Reconsider therapy in patients with no benefit after 3–4 additional doses of 100 mg

Certolizumab pegol, SC[19, 20]200 mgEvery 2 weeksFor maintenance dosing, 400 mg every 4 weeks can be considered200 mgEvery 2 weeksN/A

Dose optimization of infliximab

Dose optimization is fairly common in patients treated with infliximab[21, 22] and more studies of dose escalation have studied infliximab than other TNF inhibitors. In a systematic review of 16 observational studies of patients receiving TNF-inhibitors, 53.7% (2716/5099) of patients who received infliximab had a dose escalation (either a dose increase or interval decrease).[21] Dose optimization accomplished by an increase in the infliximab dose per infusion were more common than optimization by decreases in the interval between doses (44% vs. 8.3%),[21] despite some evidence suggesting that interval adjustment may be more effective in maintaining serum levels of infliximab[23] and achieving better clinical outcomes.[24, 25]

In the Anti–Tumor Necrosis Factor in Rheumatoid Arthritis with Concomitant Therapy (ATTRACT) trial, four different infliximab dosing regimens (with concomitant MTX) were found to be efficacious in patients with long-standing RA[3, 26, 27]: 3 or 10 mg/kg every 4 weeks or 3 or 10 mg/kg every 8 weeks. However, serum trough levels of infliximab were found to be related to clinical response, and considerable individual variability was found in trough levels within each dosage group. Pharmacokinetic modeling suggested that in order to increase trough concentrations, shortening the dose interval from 8 to 6 weeks might be more effective than raising the dose level.[23] Overall, the data from ATTRACT suggested that some patients may benefit from higher doses or more frequent dosing.[23]

Discussion of methodologic issues in evaluation of dose optimization

Early versus late dose optimization

For the purposes of this review, dose optimization that occurs early, within the first 6 months of treatment, is assumed to reflect non-response or inadequate response. Dose optimization that occurs after the first 6 months of treatment is assumed to reflect loss of response. Therapeutic failures at each time point may reflect different underlying physiologic mechanisms, and the reasons for dose changes may vary across patients. For example, some patients with initial non-response may have disease that is not primarily TNF-driven, and therefore, may not respond to this class of therapy. In fact, Takeuchi et al.[28] reported that a group of patients who had low levels of TNF prior to anti-TNF treatment did not have a dose-related response to infliximab, whereas a group of patients who had high levels of TNF before treatment responded better to higher than lower doses of infliximab.

For other patients with inadequate initial response to infliximab or other TNF inhibitors, a higher dosage may be required to develop sufficient serum infliximab concentrations for therapeutic effect.[28, 29] Another subset of patients who may respond to dose optimization are those who experience good response to the treatment, but the effect does not last for the full 8 weeks between infusions.[23] In this case, shortening the dose interval may be useful.

Some patients with initial good response to anti-TNF treatment experience diminishing response.[30, 31] Reasons for this loss of response are unclear. One possibility is that loss of response reflects normal disease fluctuation that might subside without additional treatment.[32] Another possibility is inadequate trough serum drug concentrations, which have been linked to development of antibodies to the drug over time.[33-35] Although it is unclear why only certain patients develop such antibodies, there is evidence that patients who receive concomitant MTX or other immunosuppressants are less likely than other patients to develop antibodies to TNF antagonists.[36]

Evaluation of efficacy after dose optimization

Evaluation of the evidence for efficacy after dose optimization is complicated by the fact that most studies of dose optimization are observational designs, and therefore have several limitations. Studies typically follow patients who have dose changes in the course of their treatment, and evaluate patient outcomes after the dose change. It is reasonable to think that patients who require dose optimization may have more active disease, treatment-resistant disease, or greater overall disease severity than the patients without a dose change, thus making comparisons between groups difficult to interpret. An alternative study design compares clinical status of a patient before versus after the dose change. A main concern about this type of comparison is that RA is a disease that waxes and wanes — patients may worsen, have a dose change, and get better not because of the dose change, but because the disease has natural periods of worsening and improvement. A retrospective study explored this issue in 44 patients who had infliximab dose increases (3 to 5–7 mg/kg). At a 4-month assessment, 28-joint Disease Activity Score (DAS28) and swollen joint count showed significant improvements which were maintained at later follow-up.[32] However, similar patterns of worsening followed by improvement were also observed in two control cohorts: patients treated with infliximab who had worsening response but no dose increase (= 44) and patients who were treated with etanercept (= 36). Although the observed improvement after dose escalation could be interpreted as recapturing initial response to infliximab, the similar pattern of worsening and improvement in all three groups suggests that it could also be a regression to the mean effect. That is, because doses are escalated at a time when patients are doing poorly, it is likely that they will show improvement after the dose increase, regardless of whether they have a dose change.[32]

Rigorous studies of infliximab dose optimization with a variety of methodologic approaches are needed to determine whether the perceived clinical benefit of dose optimization is actually attributable to the dose increase.

Methods

A search of literature published through to March 15, 2013 (PubMed search terms: infliximab and rheumatoid arthritis and dose; filter: search field of Title/Abstract), yielded 235 publications, from which the authors identified 12 studies that evaluated the effect of dose optimization on clinical response in infliximab-treated patients with inadequate response, nonresponse, or loss of response.[24, 25, 28, 37-45] These studies include: retrospective registry analyses; prospective observational studies; cohort studies; and randomized, controlled trials.

Results: studies of dose optimization

Table 2 includes a summary of the methods and results of each study that evaluated efficacy and/or safety of dose optimization, and each study is described below.

Table 2. Studies of dose optimization of infliximab for rheumatoid arthritis
Reason for dose optimizationStudy/referenceStudy designEfficacy resultSafety result
  1. The study protocols differed in many ways, including the types of patients included in the study, the prior treatments patients had received and the number and type of concomitant medications allowed.

  2. ACR, American College of Rheumatology; ACR20, 20% reduction in American College of Rheumatology criteria; AE, adverse event; CI, confidence interval; DAS28, 28-joint Disease Activity Score; EULAR, European League Against Rheumatism; IFX, infliximab; RA, rheumatoid arthritis; SAE, serious adverse event.

Inadequate response or nonresponse in the first 6 months of treatmentFlendrie et al.[37]Small, open-label study including patients who were partial responders (n = 26) or nonresponders (n = 21) after 14 weeks of IFX treatment and then had dose optimization (dose or frequency increase) for 24 weeksIn week-14 moderate responders, DAS28 improved significantly from 5.1 to 3.6 at week 38 in 6 patients with interval adjustment (mean dose 4.8 mg/kg/8 week), and DAS28 improved from 4.1 to 3.6 in 6 patients with dose adjustment (mean dose 7.3 mg/kg/8 week). In nonresponders, dose adjustment (mean dose 5.7 mg/kg/8 week) had less effect; DAS28 was 5.5 at baseline and 5.0 at week 38Safety data were not fully reported and not given by dose group
Durez et al.[38]Large observational study in patients treated with IFX 3 mg/kg for 22 weeks. Based on clinical judgment at week 22, 22% of patients (106/511) had a dosage increase of 100 mg per dose starting at week 30, and the rest of the patients stayed on a stable doseIn patients who had dose optimization, ACR20 increased from 34% of patients before the dose increase to almost 60% of patients at week 62. In patients who stayed at a stable dose, ACR20 increased from 60% to approximately 68%No difference in AEs in the stable versus increased dose groups
Takeuchi et al.[28]Randomized, controlled trial RISING study (n = 327). Patients started IFX at 3 mg/kg for 14 weeks (3 mg/kg at weeks 0, 2 and 6). From week 14 to week 46, regardless of response to the initial dose, patients were randomized to continue 3 mg/kg or to increase to 6 mg/kg or 10 mg/kg. A post hoc subanalysis of the 37 patients who did not have response to treatment with 3 mg/kg at week 10 was performed. Responses were compared for patients who were assigned to stay at 3 mg/kg vs. those who were assigned to dose increaseAt week 54, the 10-mg/kg group had greater improvement than the stable dose group (3 mg/kg) in mean numeric ACR response score (58.3 vs. 51.3, = 0.024), DAS28 reduction and positive EULAR response (Table 3), but not in ACR20, ACR50 and ACR70 responses. Results from the 6-mg/kg group were intermediate between the stable 3-mg/kg and 10-mg/kg groups (Table 3)In the post-hoc analysis, 90% of patients who remained at 3 mg/kg had no EULAR response. Of those who increased to 10 mg/kg, all had a moderate or good response. The 6-mg/kg group had intermediate results (57% had good or moderate response)No significant difference in the incidence of AEs in the dose groups
Loss of response after treatment of 6 months or longerSidiropoulos et al.[39]Small observational study in patients with longstanding RA who had failed multiple DMARDs. Patients were followed for up to 12 IFX infusions over a mean of 73 weeks. Clinician decisions about dose interval adjustment (within a range of 6–8 weeks) were based on DAS. Of 48 patients, 35 (73%) had interval adjustments (on average, at dose 8)DAS improved significantly from baseline to the 12th infusion for those with stable dose (4.57–2.55) and dose optimization (6.05–4.54). Disease activity was consistently greater in the dose-adjusted group than the stable dose group, including at baselineSafety data were not reported separately for patients with versus without a dose change. Overall, 10% of patients in this study discontinued treatment owing to adverse events
Edrees et al.[25]Small prospective cohort study of 55 patients with RA who had been receiving IFX 3 mg/kg every 8 weeks for approximately 2 years (range, 3–48 months). Patients with an inadequate response were randomized to 1 of 3 groups: no change (n = 13), dose increase (n = 7) or more frequent infusions (every 6 weeks; n = 11) for 16 monthsACR20 response was achieved by 18% of the group with no IFX change, 29% of the group with dose increase and 36% in the group with more frequent infusions. Joint tenderness and pain scores (Ritchie scoring index) decreased in all 3 groups, but was statistically significant only in the group with more frequent infusionsSafety data not reported
Pavelka et al.[40]Randomized study of 141 patients from the ATTRA registry who had completed 12 months of treatment with IFX 3 mg/kg. Patients were randomized to IFX 3 mg/kg (stable dose) or 5 mg/kg (increased dose) for an additional year of treatmentAt the end of the additional year of treatment, DAS28 scores had declined similarly in the 2 dose groups: by 0.67 (95% CI 0.42–0.92) in the 3-mg/kg IFX group and by 0.65 (95% CI 0.40–0.89) in the 5-mg/kg IFX groupThe groups did not differ in SAEs, serious infections, or infusion reactions; however, patients in the 5-mg/kg group had more non-serious AEs than the 3-mg/kg group (47.8% vs. 28.2%)
Inadequate response and/or non-response and loss of response: dose optimization before and after 6 months of treatmentBlom et al.[41]Observational study of patients in the Dutch Rheumatoid Arthritis Monitoring registry who had dose adjustments of IFX, adalimumab, or etanercept. Rheumatologists decided if/when to increase IFX dose or frequency, which occurred in 115/323 patients receiving IFX (median time to adjustment, 6 months), 44/368 patients receiving adalimumab (median time to adjustment, 10.5 months), and 32/420 patients receiving etanercept (median time to adjustment, 9 months)A small mean improvement in DAS28 occurred in all groups after dose optimization (statistically significant only in the etanercept group). Low disease activity (DAS28 ≤ 3.2) at 3 months after dose increase was achieved by 23.4% of the IFX group, 38.1% of the adalimumab group and 31.6% of the etanercept group; these percentages were similar 3 months laterSafety of dose increase was not evaluated in this study.
Moots et al.[44]Retrospective observational study of patients receiving IFX, adalimumab or etanercept over 18 months in normal clinical practice at 44 European centers. Dose increase occurred in 37/107 patients receiving IFX, 30/313 patients receiving adalimumab and 8/319 patients receiving etanercept. Inadequate response was the reason for the dose increase in 22% of IFX-treated patients, 8% of adalimumab-treated patients and 1% of etanercept-treated patientsEfficacy data were not reported in detail, and the number of patients with available DAS28 data both before and after dose increase was limited, but the dose increase had no measurable effect on DAS28 scoresSafety was not assessed.
van Vollenhoven et al.[24]Observational study of 420 IFX-treated patients (diagnoses were RA, juvenile-onset inflammatory arthritis, or peripheral joint predominant psoriatic arthritis) from the Stockholm registry STURE. After 3 years, 42% of patients remained on their original dose schedule; other patients had changes in dose and/or frequencyDAS28 improved from 4.12 ± 0.19 at the visit just before dose increase to 3.66 ± 0.17 at the visit following the dose increase (< 0.01). Following frequency adjustments, 63% of patients had improved persistence of response between infusions, according to chart review of physicians'/nurses' notes during the visits following the adjustmentsSafety data were not reported
Zink et al.[45]Observational study of 344 patients in the RABBIT registry treated with IFX. Mean dosage at 3 months (after the loading dose) was 3.2 mg/kg body weight/infusion and 3.3 mg/kg after 1 year. Factoring in shortened infusion intervals, the mean dosage was 4.0 mg/kg body weight every 8 weeksPatients who were treated with low dosages of up to 3 mg/kg every 8 weeks showed significantly less improvement (EULAR response) than those who were treated with higher dosagesSafety data were not reported
Westhovens et al.[42] and Rahman et al.[43]Randomized trial (Safety Trial for Rheumatoid Arthritis with Remicade Therapy [START]) of 1076 IFX-naïve patients with long-standing disease. Patients were randomized to placebo, IFX 3 mg/kg, or IFX 10 mg/kg and had infusions at weeks 0, 2, 6 and 14. At week 22 of the trial, patients who were in the 3-mg/kg group were eligible for a dose escalation in 1.5 mg/kg increments if they had inadequate response or a flare after initial responseAt week 22, the 3- and 10-mg/kg groups did not differ in levels of ACR20 response (58% and 61%, respectively) or remission (31% and 32%, respectively). Of patients in the initial 3-mg/kg group who were eligible for dose escalation at week 22 (n = 100 patients evaluable for efficacy), 59% had only 1 dose escalation (i.e., to 4.5 mg/kg). Of all patients with dose escalation, 80% responded (≥ 20% improvement from baseline in tender and swollen joint count 8 weeks after the last dose increase)The relative risk of serious infection was greater in patients receiving 10 mg/kg (3.1; 95% CI 1.2–7.9) than in patients receiving 3 mg/kg (1.0; 95% CI 0.3–3.1). In the initial 3-mg/kg group, rates of AEs and SAEs were similar for those who did and did not receive dose escalation (Table 4)

Inadequate response or nonresponse in the first 6 months of treatment

Several studies have explored the effects of dose optimization during the first 6 months of infliximab treatment. One small, open-label study investigated the effects of infliximab dose increase or interval reduction after 14 weeks of infliximab treatment in patients who were partial responders or non-responders.[37] Overall, patients who were moderate responders experienced improvement after dose escalation or interval reduction, but non-responders continued to have high disease activity.

In a much larger observational study in Belgium, patients who were MTX-refractory were treated with infliximab 3 mg/kg for 22 weeks.[38] Based on clinical judgment at week 22, 22% of patients (106/511) had a dosage increase of 100 mg per dose at week 30, and the rest of the patients stayed on a stable dose. In the group of patients with dose optimization, percentage of patients with 20% reduction in American College of Rheumatology criteria (ACR20 response) increased from 34% before the dose increase to almost 60% at week 62. In patients who stayed at a stable dose, ACR20 was relatively stable during the same period (Fig. 1). Although a regression to the mean effect cannot be ruled out, the magnitude of the effect on ACR20 response was fairly large after the dose increase, and despite greater baseline disease in the dose optimization group, the groups with and without dose optimization had similar ACR20 responses at week 62. There was no difference in adverse events in the stable versus increased dose groups.

Figure 1.

ACR20 response in patients who received infliximab 3 mg/kg until week 22 and then either had a dose increase (100 mg every 8 weeks starting from week 30) or stayed on a stable dose.[38] (Figure adapted from Durez et al. Rheumatology 2005; 44: 465–8, Fig. 1.) ACR20: 20% reduction in American College of Rheumatology criteria.

One randomized, controlled trial has provided evidence supporting a small, but statistically significant, impact of early infliximab dose optimization. In the RISING study (= 327), MTX-refractory Japanese patients started infliximab at 3 mg/kg for 14 weeks (3 mg/kg at weeks 0, 2 and 6). From week 14 to week 46, regardless of response to the initial dose, patients were randomized to continue 3 mg/kg or to increase to 6 mg/kg or 10 mg/kg.[28] At week 54, improvement was greater in the 10-mg/kg group than the stable dose group in some, but not all, measures of disease activity and function (Table 3). In the 6-mg/kg group, clinical response was intermediate between the 3 and 10-mg/kg groups. As noted in previous studies, higher trough serum levels were associated with better clinical response (measured by European League Against Rheumatism [EULAR] response), and patients achieving remission had significantly higher trough serum levels than patients without remission.[28] Overall, this randomized, controlled trial suggests that patients can benefit from increases in infliximab dose early in treatment without a significant increase in the incidence of adverse events or serious adverse events, including infections.

Table 3. Response to infliximab at three dose levels after 54 weeks of therapy in the RISING study[28]
Response measure3 mg/kg (n = 99)6 mg/kg (n = 104)10 mg/kg (n = 104)
  1. ACR-N, numeric American College of Rheumatology response; DAS-28, 28-point Disease Activity Score; EULAR, European League Against Rheumatism; HAQ, Health Assessment Questionnaire; SD, standard deviation.

  2. *< 0.05 vs. 3 mg/kg group, **< 0.01 vs. 3 mg/kg group. (Table adapted from Takeuchi et al. Mod Rheumatol 2009; 19, 478–87).

ACR-N, mean (SD)51.3 (32.1)53.8 (34.4)58.3 (31.3)*
Reduction in DAS28, mean (SD)2.30 (1.56)2.57 (1.69)2.80 (1.58)**
EULAR response, n (%)
Moderate or good78 (78.8)87 (83.7)94 (90.4)**
Good37 (37.4)52 (50.0)*52 (50.0)*
DAS28 < 2.6 (remission), n (%)25 (25.3)34 (32.7)34 (32.7)
Improvement in HAQ score, mean (SD)0.48 (0.70)0.56 (0.64)0.59 (0.63)
Rates of clinically meaningful improvement (HAQ improvement > 0.22), n (%)69 (69.7)75 (72.1)76 (73.1)

In the RISING study, Takeuchi et al.[28] also performed an interesting post hoc subanalysis exploring the effects of dose optimization in early nonresponders. Of patients who were non-responders to the 3-mg/kg dose at week 10 and remained on the same dose, 90% had no EULAR response at week 54. However, of those who were non-responders at week 10 and had a dose increase to 10 mg/kg, 100% achieved either a moderate or good response at week 54. The 6 mg/kg group had intermediate results (57% had good or moderate response).[28]

Loss of response after treatment of 6 months or longer

Few studies have explored dose optimization in patients who had initial response and then experienced loss of response after 6 months or more of treatment. In a small, uncontrolled, open-label study investigating the effects of infliximab dose optimization in patients with long-standing, active RA who had failed multiple DMARDs, patients were followed for up to 12 infliximab infusions over a mean follow-up period of 73 weeks.[39] Clinicians decided to decrease the infliximab dose interval in 35 (73%) of 48 patients. DAS28 improved significantly from baseline to the 12th infusion for those with stable dose (4.57–2.55) and dose optimization (6.05–4.54). The overall level of disease activity was consistently greater in the dose-adjusted group than the stable dose group, including a significant difference at baseline, making these findings difficult to interpret. The data suggest that dose optimization may have been helpful in reducing disease activity in these patients with very active RA; however, these patients still maintained higher levels of disease activity than patients who did not have a dose increase.[39]

In a small prospective cohort study of 55 patients with RA who had been receiving infliximab 3 mg/kg every 8 weeks for approximately 2 years, those with an inadequate response were randomized to one of three groups: no change, dose increase, or more frequent infusions for 16 months.[25] The percentage of patients with ACR20 response was 18%, 29%, and 36% in the three groups, respectively. Joint tenderness and pain scores (as measured by the Ritchie scoring index) decreased in all three groups, with a significant reduction (= 0.0349) in the group with more frequent infusions.[25]

Only one trial has randomly assigned subjects to dose escalation after an initial trial of infliximab 3 mg/kg for a period > 6 months. In a study from the Czech Republic's ATTRA registry, patients (n = 141) had completed 12 months of treatment with infliximab 3 mg/kg and had a DAS28 score > 2.6, either because initial response was inadequate or response had declined.[40] Patients were then randomized to infliximab 3 or 5 mg/kg for an additional year of treatment. At the end of the second year, DAS28 scores had declined similarly in both dose groups. The two groups did not differ in the percentages of patients with serious adverse events, serious infections or infusion reactions; however, a greater percentage of patients in the 5 mg/kg group than the 3 mg/kg group had non-serious adverse events (47.8% vs. 28.2%).[40]

Although this study was fairly small, the results suggest that, for patients who had lost response over a year of treatment, dose modification did not provide benefit at the end of one additional year of treatment. These results are in contrast to the Takeuchi et al.[28] RISING study, in which patients randomized to dose increase early in treatment had benefit over a control group with no dose change. One notable difference between the studies is that patients in the RISING study were assigned to dose groups without regard to response level at the initial 3 mg/kg dose, whereas patients in the Pavelka et al. study had all shown inadequate response or loss of response to a 3 mg/kg dose at 1 year. Another difference between the studies is that the dose increase in the Takeuchi et al. study was greater than in the Pavelka et al. study; in the Takeuchi et al.[28] study, the 10 mg/kg dose group had a significantly greater response than the 3 mg/kg dose, and the results of the 6 mg/kg dose group were intermediate. In the Pavelka et al.[40] study, the highest dose was 5 mg/kg.

Inadequate response/nonresponse and loss of response: dose optimization before and after 6 months of treatment

In several studies, dose escalations both before and after 6 months of treatment were included in the analyses, presumably reflecting a combination of patients who had initial nonresponse/inadequate response and those who had loss of response.

An observational study of patients in the Dutch Rheumatoid Arthritis Monitoring registry explored the effects of dose optimization of infliximab, adalimumab and etanercept as directed by rheumatologists in daily practice.[41] A small mean improvement in DAS28 occurred in each group following dose optimization (statistically significant only in the etanercept group). The reasons for dose optimization across the three agents were retrospectively identified as non-response, loss of response or partial response. Interestingly, a significant decrease in disease activity occurred at 3 and 6 months for non-responders and at 3 months for patients with loss of response, but no such decrease occurred for partial responders.[41]

Another retrospective observational study assessed dose escalation of infliximab, adalimumab and etanercept in 739 patients over 18 months of treatment in normal clinical practice at 44 European centers.[44] The number of patients with available DAS28 data both before and after dose increase was limited, but the dose increase did not appear to have a measurable effect on DAS28 scores.[44]

Data from the Stockholm registry STURE showed that dose optimization was common among 420 infliximab-treated patients with RA, juvenile-onset inflammatory arthritis (now adults), or peripheral joint predominant psoriatic arthritis; after 3 years, only 42% of patients remained on their original dose schedule.[24] DAS28 values improved from 4.12 ± 0.19 at the visit just before dose increase, to 3.66 ± 0.17 at the visit following the dose increase (< 0.01). After receiving more frequent infusions, 63% of patients had improved persistence of response between infusions.[24]

Data from 344 patients in the German Biologics Registry RABBIT were used to investigate how the dosage translates into clinical effectiveness.[45] Mean dosage at 3 months was 3.2 mg/kg per infusion and 3.3 mg/kg after 1 year. If we also consider shortening the infusion intervals, the mean dosages at the start and after 1 year were 4.0 mg/kg body weight every 8 weeks. Patients who were treated with low dosages of up to 3 mg/kg every 8 weeks showed significantly less improvement (EULAR response) than those who were treated with higher dosages.

The Safety Trial for Rheumatoid Arthritis with Remicade Therapy (START) study evaluated the effect of escalating doses in patients who exhibited an incomplete response to infliximab 3 mg/kg.[42] Infliximab-naïve patients (n = 1076) were randomized to placebo, infliximab 3 mg/kg, or infliximab 10 mg/kg and had infusions at weeks 0, 2, 6 and 14. At week 22, the 3 and 10 mg/kg groups did not differ in level of ACR20 response or remission. The relative risk of serious infection was found to be greater in patients receiving 10 mg/kg infliximab (3.1; 95% CI 1.2–7.9) than in patients receiving 3 mg/kg (1.0; 95% CI 0.3–3.1).[42] At or after week 22 of the trial, patients who were in the 3 mg/kg group were eligible for a dose escalation in 1.5 mg/kg increments if they had inadequate response or a flare after initial response.[43] Of the patients with dose optimization, 80% responded to the dose escalation regimen (≥ 20% improvement from baseline in tender and swollen joint count). The percentage of patients who responded to the dose escalation was similar for those who had initial non-response and those who had lost response.[43]

In an analysis of trough serum infliximab concentrations in this START subanalysis, patients who met the criteria for dose escalation generally had lower trough infliximab concentrations than those who stayed at a stable dose.[43] Rates of adverse events and serious adverse events were similar among those who did and did not receive dose escalation (Table 4). The START data suggest that reaching adequate trough serum infliximab levels is important for efficacy, but that starting patients on infliximab 10 mg/kg may not be the optimal approach for balancing safety and efficacy. The 10-mg/kg induction regimen may lead to an increased risk of infections, but a 3-mg/kg induction regimen with subsequent dose escalation (59% of patients had only one dose escalation to a dose of 4.5 mg/kg) may be well tolerated with no increase in rate of serious infections. Also of note, the proportion of patients with ACR20 response at week 22 was not substantially higher than at week 14, suggesting that patients could be evaluated for dose optimization as early as 12–16 weeks.[42, 43]

Table 4. Follow-up period and adverse events in patients with and without dose escalation of infliximab in the START trial[43]
 Patients with dose escalationa n = 109Patients without dose escalationb n = 220
  1. START, Safety Trial for Rheumatoid Arthritis with Remicade Therapy.

  2. a

    Doses ranged from infliximab 4.5 mg/kg (1 escalation of 1.5 mg/kg over starting dose) to 9.0 mg/kg (4 escalations). Overall, 59% of subjects had only one dose escalation.

  3. b

    Data from group II (infliximab 3 mg/kg) between weeks 22 and 54. (Table adapted from Rahman et al. Ann Rheum Dis 2007; 66, 1233–8, Table 3).

Weeks of follow-up (mean)31.731.9
Weeks of treatment (mean)30.631.0
Patients with adverse events, n (%)
Patients with ≥ 1 adverse event(s)77 (70.6)160 (72.7)
Patients with ≥ 1 serious adverse event(s)14 (12.8)19 (8.6)
Patients with infections39 (36)80 (36)
Patients with serious infections2 (1.8)5 (2.3)

Discussion

In the last two decades, it has become clearer that RA disease is heterogeneous, varying across patients in terms of disease course, rate of progression and extent of joint damage.[46, 47] The individual variation in disease course means that treatment should be individualized, that one treatment strategy may not work for every patient, and that each patient may have different treatment needs at different times. Variability in disease course also means that close monitoring is important in order to have treatment that is responsive to the patient's current status. Several studies have shown that intensive outpatient strategies involving tight control and frequent monitoring of disease activity, with treatments adjusted to meet therapeutic goals (e.g., DAS28 remission), can produce better outcomes than traditional strategies.[9, 10, 48] However, it may be important for clinicians to avoid basing their treatment decision on a single indicator. For example, in a patient with join pain, if the disease activity score is mainly driven by the number of painful joints and the patient global assessment, but the patient has normal erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) and no swollen joints, then inflammation may not be the cause of the patient's pain, and a dose optimization strategy may be ineffective.

Many patients who receive TNF inhibitors show an adequate response to initial treatment. However, some patients fail to show initial response, have an inadequate response, or lose response over time. Studies of dose optimization in these patients are not entirely consistent, but tend to show a pattern of improvement after a dose increase. The effect is small in some studies, and difficult to distinguish from a regression to the mean effect in some trials. Two randomized trials conducted to study the effect of infliximab dose increase showed different results; a dose increase early in treatment led to better response,[28] but a dose increase after 1 year of treatment did not lead to greater improvement at year 2 when compared with patients who did not receive a dose increase.[40]

Dose optimization with TNF inhibitors should balance the potential for improved efficacy with the potential risks associated with higher doses. The studies reviewed here have mixed results; some report an increase in adverse events or serious adverse events for patients who receive higher doses, and other studies do not. Several of the studies have study designs or other methodologic features that make evaluation of the association between dosing and adverse events difficult to evaluate.

Another issue of practical concern is the economic impact of dose escalation, which is an important factor in clinical decisions about treatment, and is affected by reimbursement guidelines that vary considerably across countries. Several recent papers have explored the economic impact of dose optimization of infliximab and other anti-TNF agents.[49-51]

In future studies of dose optimization, it would be helpful to more fully explore the relationships between drug levels, treatment response and presence of anti-drug antibodies. Serum trough levels of TNF inhibitors are related to their clinical effectiveness, and patients who develop antidrug antibodies usually have lower drug levels and poorer response to treatment. Some evidence suggests that concomitant immunosuppressant treatment (e.g., with MTX) may reduce the development of antidrug antibodies,[35] and drug concentrations can be amplified by dose optimization, but few studies have carefully monitored over time the relationships between dosing, clinical response and development of antibodies.

Conclusions

Overall, the available data from studies of dose optimization suggest that some patients benefit from it. More research is needed to identify specific subpopulations that might benefit most from dose optimization and develop clear guidelines to aid clinicians in decisions about when dose changes are warranted.

Acknowledgements

Merck Sharp & Dohme Corporation, Whitehouse Station, New Jersey, USA, sponsored and supported development of this manuscript. The authors thank Nathan Vastesaeger, MD, and Denesh Kumar Chitkara, MD, who are employees of Merck, for helpful comments on a draft version of this manuscript. The authors determined the final content of the manuscript. No payments were made to the authors for the writing of this article. Medical writing assistance was provided by Jacqueline M. Mahon, MA, of Synergy Medical Education, Conshohocken, PA, and Ellen Stoltzfus, PhD, of JK Associates, Inc., Conshohocken, PA. This assistance was funded by Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, NJ. RA has served on an advisory board for Merck Sharp & Dohme and has been a speaker for Schering-Plough/Merck Sharp & Dohme. FvdB has served on an advisory board for Merck Sharp & Dohme and has been a speaker for Schering-Plough/Merck Sharp & Dohme/Janssen-Centocor.

Authors' contributions

Both authors are responsible for the content of this manuscript. The authors together conceptualized the manuscript content, reviewed multiple drafts and approved the final version for submission.

Ancillary