Background/Purpose: The long-term cost-effectiveness of triple therapy (methotrexate [MTX], sulfasalazine [SSZ], hydroxychloroquine [HCQ]) disease modifying anti-rheumatic drugs (DMARDs) compared to a combination of MTX and anti-TNF therapy in patients with early rheumatoid arthritis (RA) is unknown. In this study, we used patient-level data from the Treatment of Early Aggressive RA (TEAR) trial and the National Data Bank for Rheumatic Diseases (NDB) to evaluate the cost-effectiveness of immediate combination therapy (with biologic or non-biologic DMARDs) versus stepping up to combination therapy at 6 months if disease activity persists despite MTX monotherapy.

Methods: We developed a Markov simulation model to estimate quality-adjusted life years (QALYs) and costs associated with the treatment strategies examined in the TEAR trial.  We evaluated four strategies: immediate triple (IT), immediate etanercept (IE), step-up triple (ST), and step-up etanercept (SE).  The step-up strategies involved switching those with persistent disease activity (DAS28 >= 3.2) from MTX monotherapy to MTX plus either etanercept or triple therapy at 6 months.  The simulation model extends the 2-year trial to the life-time horizon, using parameters taken from longitudinal NDB data for therapy discontinuation rates, HAQ transitions, and DAS28-HAQ-QALY mappings as well as the published literature for direct and indirect cost estimates.  Annual discontinuation rates of triple therapy and etanercept from the NDB were estimated to be 22% and 10%, respectively.   Those who discontinued were assumed to continue to receive methotrexate.  Markov health states were defined by DAS28. DAS28 score transitions were obtained directly from individual patients in the trial. Death was modeled as an additional state with background mortality estimated from the 2007 US Life Tables. HAQ scores were used as a secondary variable to estimate QALYs, RA-specific mortality, and direct and indirect costs (e.g., due to productivity loss). We assumed a 3-month cycle length, and discounted both costs and effectiveness by 3% annually.

Results: The etanercept strategies (SE and IE) were more costly than the triple strategies (ST and IT) mainly due to treatment costs [Table].  The lifetime benefits from IT, ST and SE were numerically similar (within 0.06 QALYs).  Although IE was more effective than IT, the incremental cost-effectiveness ratio (ICER) of IE relative to IT was $837,100/QALY.  These results were robust to parametric sensitivity analyses.

Conclusion: We used patient-level data from the TEAR trial, and then projected their lifetime costs and benefits using the NDB.  The benefits from all strategies were comparable, but biologics strategies were almost twice more expensive than triple strategies, producing ICERs greater than what most healthcare settings find acceptable. 

Table.  Cost-Effectiveness Analysis of TEAR Strategies
Treatment Strategies	Cost ($)	Effectiveness (QALY)	Incremental Cost ($)*	Incremental Effectiveness (QALY)*	ICER ($/QALY)
Immediate Triple (IT)	152,400	9.991			Reference Strategy
Step-up Triple (ST)	154,900	9.928	2500	-0.063	Excluded (dominated by IT, more effective)
Step-up Etanercept (SE)	269,500	9.929	117,100	-0.062	Excluded (dominated by IT, more effective)
Immediate Etanercept (IE)	338,100	10.213	185,700	0.222	837,100
*Compared to IT

---

« Back to 2013 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/cost-effectiveness-analysis-of-triple-therapy-versus-etanercept-plus-methotrexate-in-early-aggressive-rheumatoid-arthritis-analysis-based-on-the-tear-trial/