BACKGROUND

Crohn's disease (CD) and ulcerative colitis (UC) are chronic inflammatory bowel diseases (IBDs). Intermittent courses of corticosteroids and chronic immunosuppression with anti–tumor necrosis factor α (anti-TNF) drugs represent competing treatment strategies for IBD, each with potential benefits and harms.

AIMS

(1) To assess heterogeneity in patients' stated preferences for health states and medication-related risks relevant to IBD; (2) to compare the mortality, infection, and surgery risk with prolonged corticosteroids use vs anti-TNF drugs in patients with IBD; and (3) to compare quality of life with prolonged corticosteroids use vs anti-TNF drugs in patients with IBD.

METHODS

Aim 1: A stated-preference survey was administered to CD patients asking them to choose between pairs of constructed medical therapies for moderately active CD. Treatment options were characterized by differing levels of time with active disease symptoms; severity of symptoms; duration of therapy with steroids; and risks of serious infection, cancer, and need for surgery. Latent class choice models identified groups of patients with similar treatment outcome preferences. We converted reference weight estimates derived from stated preference surveys to a new measure of utility—remission time equivalents (RTEs)—to provide a metric for aim 3. For aims 2 and 3, we conducted a retrospective cohort study of Medicaid and Medicare beneficiaries with IBD in the United States from 2006 to 2013. The primary exposure variable was receipt of >3000 mg of prednisone within 12 months or new initiation of anti-TNF therapy. In aim 2, the primary outcome was all-cause mortality. We used marginal structural models to determine odds ratios (ORs) and 95% CIs for anti-TNF use relative to corticosteroids for all outcomes. In aim 3, after matching on propensity scores, we compared mean total RTEs in the first 12 months of follow-up between treatment groups using a paired test. We built a Markov model to mimic the results of the cohort study. We used the Markov model to perform many additional sensitivity analyses.

RESULTS

Latent class analysis demonstrated 3 distinct groups of patients whose choices were strongly influenced by (1) duration of active symptoms (61%); (2) duration of steroid use (25%); or (3) avoidance of risks of cancer, infection, or surgery (14%). In aim 2, relative to treatment with prolonged corticosteroids the risk of death was statistically significantly lower in patients treated with anti-TNF therapy for CD (odds ratio [OR], 0.78; 95% CI, 0.65-0.93) but not statistically significantly reduced for UC (OR, 0.87; 95% CI, 0.63-1.22). Among patients with CD, anti-TNF therapy was also associated with lower rates of Major adverse cardiovascular event (MACE) (OR, 0.68; 95% CI, 0.55-0.85) and hip fracture (OR, 0.54; 95% CI, 0.34-0.83). The risk of serious infection did not differ by treatment for either disease, but the risk of emergency surgery was higher in the anti-TNF-treated patients with UC (OR, 2.18; 95% CI, 1.37-3.46), likely due to incompletely adjusted confounding by indication given that anti-TNF therapy is often used as a final attempted therapy before surgery. Applying RTEs to the cohort study demonstrated that treatment with anti-TNF therapy yielded greater quality of life in the first 12 months after cohort entry (difference = 0.80 RTEs; 95% CI, 0.53-1.07). The Markov model analyses demonstrated that the estimated benefit of anti-TNF therapy was not sensitive to transition probability or RTE estimates.

CONCLUSIONS

Compared with corticosteroids, anti-TNF use was associated with reduced mortality in patients with CD. Treatment with anti-TNF therapy also yielded greater quality of life for all subgroups of patients with CD despite substantial heterogeneity among CD patients' preferences for medication efficacy and potential harms.