PTH-078 Thiopurine maintenance therapy for ibd: which is the best method to measure medication adherence?

    loading  Checking for direct PDF access through Ovid



For the majority of patients with IBD long-term therapy is required to maintain remission; yet 30%–45% of patients do not adhere to their IBD medication. Medication adherence can be assessed with prescription refill rates, biological measures (metabolites, trough levels, etc) and patient self-report tools. There is currently no accepted gold standard and the feasibility and utility of different adherence assessment tools in the routine outpatient clinic setting has not been fully examined. The aim of this service improvement project was to test the acceptability of self-report tools assessing thiopurine adherence in the IBD clinic and to correlate the results with thioguanine-nucleotide (TGN) levels.


Consecutive outpatients on thiopurine maintenance therapy for IBD for >3 months were recruited from clinic. Patients self-reported adherence using a visual analogue scale (VAS), the validated Morisky adherence tool (MOR) and the validated Medication Adherence Report Scale (MARS). TGN levels were classed as complete non-adherence (<100 and MMP low), partial adherence (TGN 100–235 and MMP low) or full adherence (>235 or MMP high). Correlation analysis was performed using Pearson tests.


Of 100 approached patients none refused participation and TGN levels were available for 69. These included 38 women. Diagnoses were Crohn’s disease in 27, ulcerative colitis in 41 and IBD-U in 1 cases. Concomitant therapy included 5-ASA (25 cases), anti-TNF (13 cases) and Vedolizumab (2 cases). The proportion of adherent patients was according to the relevant report tool 71% (TGN), 87% (VAS), 87% (Morisky) and 77% (MARS). VAS (Pearson 0.315; p=0.005) and Morisky (Pearson −0.363; p=0.001) correlated moderately with TGN, but MARS (Pearson 0.09; p=0.39) did not. The 7 patients, who were non-adherent by TGN were detected by VAS in 3, Morisky in 6 and MARS in 3 cases. However, patients showing non-adherence according to self-report tools had normal TGN levels in 6 of 10 cases for VAS, 10 of 26 for Morisky and 4 of 15 for MARS.


Self-report tools provided a patient friendly and inexpensive way of assessing adherence, but correlation with TGN levels was only moderate. While providing a more objective assessment TGN levels are problematic for routine use in all patients. TGN require a more invasive and expensive approach. Furthermore TGN cannot detect “white coat adherence” (patients take medication only around appointments), which is the most likely explanation for normal TGN levels in patients reporting to be poorly adherent. Neither TGN levels nor self-report tools can be seen as the gold standard at present.

Disclosure of Interest

None Declared

Related Topics

    loading  Loading Related Articles