Dr. Solomon receives royalties from UpToDate. He has received support through research grants to Brigham and Women's Hospital from Amgen, Lilly, Pfizer, and the Consortium of Rheumatology Researchers of North America (CORRONA) registry and has served in unpaid roles in trials funded by Pfizer, Novartis, Lilly, and Bristol-Myers Squibb.
Brigham and Women's Hospital, Boston, Massachusetts
The treatment armamentarium for rheumatoid arthritis (RA) has grown substantially over the last 15 years since the development of targeted biologic and nonbiologic disease-modifying antirheumatic drugs (DMARDs). These drugs have broadened the treatment possibilities and changed how rheumatic disease experts approach the clinical management of RA. The goal of reducing disease activity to very low levels (or remission) is now realistic, and emerging evidence suggests that treating to achieve these targets enhances long-term structural and quality of life outcomes ([1-7]). Consequently, “treat to target” (TTT) has become an attractive concept in the clinical management of RA. TTT is generally defined as a treatment strategy in which the clinician treats the patient aggressively enough to reach and maintain explicitly specified and sequentially measured goals, such as remission or low disease activity. TTT is proactive, has a clear end point (the “target”), and can be operationalized as a specific treatment algorithm, simplifying the multitude of complex medication sequences that can be used to treat active RA. The emerging TTT paradigm is supported by findings from many randomized controlled clinical trials in the last decade, not designed as TTT strategy trials, that suggest the benefits of early aggressive treatment approaches ([5, 8, 9]).
While TTT has many potential benefits, the rheumatology community needs to critically appraise its value in the treatment of RA. Put another way, is TTT a proven paradigm or a hypothesis requiring more complete testing? In this review, we first describe how TTT has been defined in RA and what data support its use. Second, we examine the conceptual roots of TTT, assessing how it has been used in conditions other than rheumatic diseases. Third, we examine current DMARD use patterns and barriers to TTT in clinical practice. Fourth, we discuss data from the patients' perspective relevant to TTT and set out a research agenda to address identified gaps in knowledge.
Treat to target in rheumatoid arthritis
Over the past 10–15 years, several randomized controlled clinical trials have demonstrated that a TTT strategy can achieve superior clinical outcomes compared with usual care. Studies that have provided evidence for TTT can be divided into randomized strategy trials, assessing the efficacy of treating to a specific target versus routine care in the comparator arm, and treatment target trials in which all treatment arms have a defined target but different treatment strategies to reach the target are compared. All TTT trials have included relatively frequent assessment with recommendations for intensifying treatment when patients have not reached the target. These trials (listed in Table 1) have been primarily conducted in Western Europe ([1-7, 10]). The number of subjects included in TTT trials ranges from 96 to 508. Some of the trials required subjects in the TTT arm to be treated according to specific treatment algorithms, and others allowed treating physicians to decide on treatments but required a specific disease activity goal. Almost all of the trials randomized at the subject level. Treating providers were not blinded with regard to assignment group in most studies, but many of the trials used blinded assessors. Duration of followup ranged from 6 to 36 months, and few trials accounted for the clustering of subjects within practices.
Table 1. Selected treat-to-target randomized controlled trials in rheumatoid arthritis*
Strategy trials refer to trials in which one group was treated to target and the other was not, and treatment target trials refer to trials in which all groups were treated to the same target but the strategies used to attempt to achieve the target differed. TTT = treat to target; TICORA = Tight Control of Rheumatoid Arthritis; NHS = National Health Services; EULAR = European League Against Rheumatism; DAS28-ESR = Disease Activity Score in 28 joints using the erythrocyte sedimentation rate; CAMERA = Computer Assisted Management in Early Rheumatoid Arthritis; SJC = swollen joint count; TJC = tender joint count; PGA = physician's global assessment; CRP = C-reactive protein; ULN = upper limit of normal; NS = not significant; FIN-RACo = Finnish Rheumatoid Arthritis Combination Therapy; BeST = Behandelstrategieën voor Reumatoide Artritis (Treatment Strategies for Rheumatoid Arthritis).
aResults for the studies by Grigor et al, Verstappen et al, Fransen et al, and Saunders et al are the percentages of subjects in the TTT and control groups who met the criteria. Results for the study by Symmons et al are the mean Health Assessment Questionnaire (HAQ) scores. Results for the study by Mottonen et al are the percentages of subjects receiving triple therapy and those receiving monotherapy (control) who met the criteria, and results for the study by Goekoop-Ruiterman et al are the percentages of subjects receiving initial combination therapy with infliximab and those receiving initial monotherapy (control) who met the criteria.
bAmerican College of Rheumatology (ACR) 1981 remission criteria.
20 rheumatology departments in The Netherlands (n = 508)
DAS44-ESR <2.4 (low disease activity) at year 2
Yes; 12 weeks
Low disease activity
The treatment targets varied across trials, and the rates of attaining the targets in the intervention arms ranged from 31% to 82%. Rates of reaching the target were higher in the intervention arm than in the control arm in all but one study. The safety of TTT was comparable to that of the non-TTT arms, with no greater rate of withdrawal due to adverse events. Several studies, but not all, noted reduced progression in radiographic measures. Almost no information is available regarding the cost of a TTT strategy ().
Data on the prognostic importance of consistent control of disease activity gave birth to the TTT strategy. Achieving optimal outcomes and aiming for targets using treatment strategies with maximum benefit and minimal harm was the biggest motivation for developing recommendations for the treatment of RA (). An international task force recently issued recommendations about TTT (). While these recommendations are not uniformly accepted, several aspects of these TTT recommendations and principles are important to highlight (see Table 2). Remission is specified as the primary target, but the recommendations note that low disease activity may be an appropriate alternative target. The American College of Rheumatology (ACR) RA treatment guidelines also point out the importance of these treatment goals (). It is further noted that the choice of the disease activity measure and target may be influenced by comorbidities and drug toxicities, and that patients must be appropriately informed about the treatment target. Furthermore, the TTT recommendations embed several important principles, including that RA treatment should be based on a model of shared decision making. They also appropriately note that many aspects of TTT are based on limited evidence.
Table 2. Selected European League Against Rheumatism recommendations and principles of treat to target (TTT) in rheumatoid arthritis (RA)*
The primary target for the treatment of RA should be a state of clinical remission.
While remission should be a clear target, based on available evidence low disease activity may be an acceptable alternative therapeutic goal, particularly in patients with established disease of long duration.
The patient should be appropriately informed about the treatment target, the strategy planned to reach this target under the supervision of the rheumatologist, and the risks and benefits of a TTT approach. The patient should be involved in refining the target to ensure that it is congruent with the patient's values and preferences.
The treatment of RA must be based on a shared decision between patient and rheumatologist.
The primary goal of treating the patient with RA is to maximize long-term health-related quality of life through control of symptoms, prevention of structural damage, and normalization of function and social participation.
Thus, the recommendations are both subtle and complex, requiring both that providers elicit the patient's treatment goals and preferences and that providers pursue a standardized treatment algorithm with an objective treatment target. While the algorithm will result in treatment that is no more aggressive than is typical for most patients and providers, some patients will experience an escalation in treatment beyond what would typically be prescribed.
Treat to target outside of rheumatology
TTT has become a popular concept in the medical management of several common chronic conditions, including diabetes mellitus (DM), hypertension, and hyperlipidemia. In its early formulation, TTT was used in the care of DM to design trials that focused on a glycosylated hemoglobin (HbA1c) target, as opposed to specific treatment algorithms or combinations (). The impressive reductions in long-term DM-related complications and overall mortality seen in the Diabetes Control and Complications Trial (of type 1 DM) and the United Kingdom Prospective Diabetes Study (of type 2 DM) solidified consensus around threshold target-based therapy ([15, 16]). The concept spread rapidly to the hypertension and lipid arenas, where similar studies were undertaken with explicit blood pressure and low-density lipoprotein (LDL) goals, respectively. Strategy trials, using a TTT approach, were undertaken and often utilized nonphysician providers and explicit algorithms for medication use ([17, 18]).
The substantial long-term outcome data from trials focused on lowering LDL, blood pressure, or HbA1c lay a strong foundation for TTT. For example, a large meta-analysis of 14 statin trials showed that reaching target LDL levels reduced mortality by 12% (). As a result of these data, the National Cholesterol Education Program Adult Treatment Panel recommendations rest upon Level I evidence (). That said, in the context of a randomized controlled strategy trial, more aggressive treatment targets for HbA1c among patients with DM and cardiovascular disease produced worse outcomes in the arm with a more aggressive target ().
The success of TTT in other chronic conditions serves as an important motivator for a TTT strategy in RA. RA is similar to these conditions in that all are chronic conditions with effective treatments, and combination therapy is frequently required to control the disease (Table 3).
Table 3. Comparison of characteristics of rheumatoid arthritis with characteristics of other chronic conditions in which treat to target is an accepted paradigm*
Other chronic conditions (diabetes mellitus, hypertension, and hyperlipidemia)
DAS = Disease Activity Score; CDAI = Clinical Disease Activity Index; RAPID = Routine Assessment of Patient Index Data.
Generally safe; little monitoring required for most medications except for subcutaneous insulin, which requires daily monitoring
Evidence for tight control
Relatively weak evidence of long-term benefits
Strong evidence of long-term benefits, some evidence of risks
While the similarities between RA and these conditions in which TTT has been used are important, several differences are also notable (see Table 3). First, RA is generally symptomatic, and consequently, patients can report their disease activity using validated self-report measures. In stark contrast, DM, hypertension, and hyperlipidemia are often, but not always, asymptomatic. The lack of symptoms can contribute to “clinical inertia” in which providers and patients become reluctant to change therapies despite imperfect disease control as evidenced by numerical targets ().
Second, treatments for RA require substantial disease monitoring for potential harm, including the risk of infections, heart failure, or liver disease. While the absolute risk of these treatments is not likely different from that of many treatments for other chronic diseases, they are perceived by most patients (and many providers) as “dangerous” drugs because of their effects on the immune system and black box warnings related to cancer and infection. In addition, most RA treatments have accompanying recommendations for laboratory monitoring, which may add to the perception of risk. Such perceptions may create patient concerns about increasing doses or adding or changing treatments.
Third, the evidence linking tight disease control with long-term improved outcomes is less robust for RA than it is for DM, hypertension, or hyperlipidemia. Several studies suggest that tight long-term control in RA is more likely than routine care to help a patient achieve clinical remission and minimize radiographic damage, but efficacy varies among patients, and many still incur radiographic damage over time ([1-6]). In contrast, the association between achieving target disease control in DM, hypertension, and hyperlipidemia and reduced long-term morbidity has been studied in many thousands of patients over several decades. It is noteworthy that, along with beneficial effects, tight control of hyperglycemia and hypertension has also been associated with serious risks, such as symptomatic hypoglycemia or hypotension, and even increased mortality in some subgroups with multiple comorbidities ().
Barriers to treat to target in rheumatology
If we accept the benefits of a TTT paradigm in RA, many formidable challenges limit incorporating it into typical practice. First, it appears that many patients with RA do not receive even a DMARD. Population-based studies (not those performed in rheumatology practice) demonstrate that 35–60% of patients with at least 2 diagnoses of RA do not have a record of filling prescriptions for DMARDs (). It is likely that some of these patients do not have RA, have very mild disease, or have contraindications to DMARDs. However, the sizable proportion of RA patients not receiving DMARDs suggests that widespread deployment of a TTT strategy will be challenging.
Second, few nonrheumatologists are comfortable managing DMARDs, and relatively few patients with RA have easy access to rheumatologists. A recent ACR workforce study demonstrated that 83% of urban areas in the US with populations between 10,000 and 50,000 have no rheumatologist; in those regions, the median distance to the nearest rheumatologist is 159 miles (). We also know that the strongest predictor of patients with RA receiving DMARDs is a visit with a rheumatologist ([25-27]). Thus, without better access to rheumatologists, much of the RA population will likely not have a chance of being treated with a TTT strategy.
Third, even if patients can access a rheumatologist, many patients may not want to pursue a TTT strategy. As noted in Table 2, enlisting patients as partners in TTT is an important principle. However, it appears that a sizable portion of patients are reluctant to change treatments if they feel “okay” despite having active arthritis. In one study of more than 6,000 subjects from the National Data Bank on Rheumatic Diseases, 77% indicated they were satisfied with their medications, yet 71% of these satisfied respondents had moderate or greater disease activity as assessed by their Patient Activity Scale and Health Assessment Questionnaire scores. These observations demonstrate discordance between patients' treatment preferences and their perceived pain and function (). Similar work from other researchers also demonstrates discordance between rheumatologists' ratings of disease activity and their patients' ratings (). A possible explanation for this discrepancy may be that patients observe improvement in their disease state, even though they have not reached the desired target for disease activity. In addition, patients may not be convinced that changing treatments will actually improve outcomes.
Fourth, TTT requires frequent visits and the use of structured RA disease activity measures, 2 potential impediments for rheumatologists with busy practices. While return visits for patients who have not reached target disease activity should be brief, frequent visits may be difficult for patients. Consistent use of structured disease activity measures (such as the Clinical Disease Activity Index, Simplified Disease Activity Index, Disease Activity Score in 28 joints, and Routine Assessment of Patient Index Data [RAPID]) is required for a TTT strategy, but it is unclear whether most rheumatology practices use them consistently.
Finally, since TTT requires rapid treatment escalation in the face of ongoing active disease, medications must be available without long delay. As any US-based provider knows, gaining approval for expensive treatments in RA is not quick and is often burdensome. This barrier, plus expensive patient copayments, is not likely to lessen as the pressure to reduce health care costs continues.
Patient perspectives on treat to target
One of the principles for TTT described by the international task force includes engaging patients in a discussion about their goals for treatment. Relatively little has been written about RA patients' understanding or attitudes regarding TTT. However, studies of other chronic conditions for which TTT has been used present some information regarding patient perspectives.
It is clear that despite well-documented campaigns to spread the provision of targeted diabetes therapy to achieve goal HbA1c levels, patient buy-in and adherence are still suboptimal (). A host of patient-level barriers exist, including medication ease of dosing, cost, inconvenience, lack of disease symptoms, medication side effects, or disinterest in the frequent monitoring necessary for glycemic control. Similar issues exist for both hypertension and hyperlipidemia.
While we do not yet have studies specifically focusing on RA patients' attitudes toward TTT, several studies suggest that implementing and adhering to TTT strategies in clinical practice will be challenging. Findings from the National Data Bank on Rheumatic Diseases demonstrated that many patients are satisfied with their current medications despite having levels of disease activity that would warrant escalation of care according to TTT algorithms. Moreover, patients may be reluctant to change treatment regimens not only because of the fear of side effects associated with new medications, but also because of a fear of losing control of their disease. Recent data suggest that high disease activity (as indicated by RAPID-4 scores) is predictive of future escalation only in patients who also report that their illness has a significant physical and/or emotional impact on their quality of life (). Thus, patients who have adapted to their disease may not have changes on patient-reported outcome measures and may be unlikely to be willing to escalate treatment, regardless of their disease activity score (). Data from TTT trials related to improvements in outcomes that patients can relate to, such as quality of life, would likely help patients understand the value of TTT.
TTT recommendations specify the importance of adhering to a shared decision-making model. Given the structure inherent in the TTT approach, this model requires that patients are fully informed of the specific algorithms to be used and that they agree to an increased burden of monitoring and likely escalations of care (even, in some cases, when they do not feel that additional medications are required). Shared decision making may be easier to implement in TTT strategies using specific targets as opposed to those requiring the use of specific medications, since the former enables physicians to incorporate their patients' treatment preferences. There has been some effort to make TTT recommendations more patient friendly (); this effort needs to continue.
It is important to consider the patient perspective in other diseases for which TTT has been used. Some literature suggests that while patient function may improve when a TTT strategy is used, other data suggest that the burden of treatment (such as checking blood sugar, more needle sticks) can increase depression scores ([34, 35]). In the setting of RA, the targets for treatment mix both objective scores (joint counts and markers of inflammation) with the patient experience (patient global assessment). Thus, TTT in RA relies on treating toward a physiologic target, but the patient must experience the target as a stepping stone to enhanced quality of life. Patients must be educated adequately to fully endorse the target in a TTT approach.
In conclusion, TTT in RA faces many challenges limiting its widespread acceptance. Some of these challenges are scientific due to a relatively sparse evidence base. We have outlined some of the major questions facing TTT (see Figure 1). These can be fit into several categories of research: biologic, clinical, and health services. These topics may serve the rheumatology community well as areas for research proposals.
Other barriers include potential conceptual mismatch: patients know how their arthritis affects their body and what they want from their treatments; their goals may not align with an objective target of low disease activity or remission. Further, some patients may be too fearful of the potential risks of aggressive therapy to engage in TTT. Moreover, limited access to rheumatic disease expertise limits the use of DMARDs in the US and certainly will limit the dissemination of TTT. We believe that there is a continued need for testing and refining many of the concepts underpinning TTT in RA. We look forward to a robust research effort in response to the important potential posed by TTT. Clearly, substantial evidence indicates that use of a TTT strategy in RA holds promise. However, many aspects of TTT need more data to push it from a hypothesis to a fully proven treatment strategy.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Solomon had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.