Dr S. A. Riley, Department of Gastroenterology, Northern General Hospital, Herries Road, Sheffield S7 5AU, UK. E-mail: email@example.com
Background: Non-compliance with maintenance mesalazine therapy may be a risk factor for relapse in inflammatory bowel disease, but the prevalence and determinants of non-compliance are unknown.
Aim: To study the prevalence and determinants of non-compliance in patients with inflammatory bowel disease.
Methods: Out-patients receiving delayed-release mesalazine were studied. Compliance was determined by direct enquiry and by analysis of urine samples for 5-aminosalicylic acid/N-acetyl-5-aminosalicylic acid. Potential determinants of compliance were assessed.
Results: Ninety-eight patients were studied. Forty-two patients (43%) reported taking < 80% of their prescribed dose. Logistic regression revealed the independent predictors of non-compliance to be three-times daily dosing [odds ratio (OR), 3.1; 95% confidence interval (CI), 1.8–8.4] and full-time employment (OR, 2.7; 95% CI, 1.1–6.9). Urine from 12 patients (12%) contained no detectable 5-aminosalicylic acid/N-acetyl-5-aminosalicylic acid, and 18 patients (18%) had levels below those expected. Depression was the only independent predictor of complete non-compliance (OR, 10.5; 95% CI, 1.8–79.0), and three-times daily dosing was the only independent predictor of partial non-compliance (OR, 3.7; 95% CI, 1.8–8.9). Self-reporting correctly identified 66% of patients judged to be non-compliant on urinary drug measurement.
Conclusions: Non-compliance with maintenance mesalazine therapy is common in patients with inflammatory bowel disease. Three-times daily dosing and full-time employment are predictors of partial non-compliance, whilst depression is associated with complete non-compliance. Self-reporting detects most non-compliant patients.
Delayed-release mesalazine has a well-established role in the management of patients with inflammatory bowel disease.1 It is of some benefit in active disease, but is of particular value when taken regularly, as it reduces the relapse rates in patients with ulcerative colitis1–4 and possibly also in those with Crohn's disease.5,6 Furthermore, recent studies have suggested that mesalazine-based drugs may reduce the increased risk of colorectal cancer in some patients with inflammatory bowel disease.7–9 For these reasons, most patients are encouraged to take life-long maintenance treatment.
Unfortunately, non-compliance with maintenance medication in patients with inflammatory bowel disease is common10–12 and, although the data are limited, one recent study has suggested that such patients are at increased risk of relapse.13 The reasons for non-compliance are often complex. A variety of patient-, disease-, treatment- and doctor-related factors have been implicated in a variety of disease states,14–20 but there has been no systematic study in patients with inflammatory bowel disease.
Furthermore, compliance itself is difficult to study. Patient self-reporting, physician estimation, pill counts, the use of pharmacy records and even electronic monitoring of pill dispensing have been employed, but may not truly reflect drug intake.20–24 The determination of a drug or its metabolite in biological fluids is a more objective measure of consumption,20,21,25 but this has not been studied previously as a measure of compliance with delayed-release mesalazine in patients with inflammatory bowel disease.
The aims of the present study were to investigate the use of urinary drug excretion as an objective measure of compliance, and to study the prevalence and determinants of compliance with maintenance mesalazine in a cohort of patients with inflammatory bowel disease using both patient self-reporting and urinary drug testing.
Urinary drug excretion in healthy volunteers
In order to establish the range of urinary drug excretion in spot samples during various mesalazine dosage regimens and withdrawal from those regimens, 10 healthy volunteers were recruited. Subjects were selected by their willingness to be strictly compliant. Each was asked to take three commonly used regimens of delayed-release mesalazine (Asacol, SmithKline Beecham, Uxbridge, UK): 400 mg twice daily, 800 mg twice daily and 800 mg three times daily. Each regimen was taken for seven consecutive days with a minimum 7-day drug-free interval. Doses were taken at 09.00, 14.00 and 21.00 h as dictated by the regimen. The need for strict compliance was stressed and a tablet count was performed to corroborate adherence to each regimen.
Urine samples were collected on days 7–10. On days 7, 8 and 9, spot samples of urine were collected at hourly intervals during the normal working day (09.00, 10.00, 11.00, 12.00, 14.00, 15.00, 16.00 and 17.00 h). On day 10, a single spot urine sample was collected at 09.00 h. Urine samples were frozen at − 20 °C prior to analysis.
Patient compliance study
Patients were recruited from the out-patient clinics at the Northern General and Royal Hallamshire Hospitals in Sheffield, UK. Patients with ulcerative colitis, Crohn's disease or indeterminate colitis were eligible for the study if they were receiving maintenance therapy with delayed-release mesalazine (Asacol). Patients in whom the dose of mesalazine had been changed in the last 4 weeks and those using other mesalazine-containing formulations or drugs known to alter gastrointestinal tract pH were excluded. Patients were unaware of the study prior to clinic attendance. All patients provided written informed consent and the protocol was approved by the local research ethics committees.
Patients underwent a structured interview detailing the disease characteristics and medication use. Each was asked to gauge his or her own compliance (0–100%). Patients then completed a number of validated questionnaires assessing: (i) disease activity (Simple Clinical Colitis Activity Index for ulcerative colitis26 or the Harvey–Bradshaw Index for Crohn's disease27); (ii) quality of life (Short Inflammatory Bowel Disease Questionnaire28); (iii) the doctor–patient relationship (Stanford Trust in Physician Scale29); and (iv) psychiatric morbidity (Hospital Anxiety and Depression Scale30).
Finally, patients were asked to provide a spot sample of urine. The time was noted and the sample was frozen at − 20 °C prior to analysis.
Urinary 5-aminosalicylic acid (5-ASA) and N-acetyl-5-ASA were measured by high-performance liquid chromatography, using a previously described method.31 Calibration curves were linear over the range 0.1–1000 µg/mL, and the limit of detection for both 5-ASA and N-acetyl-5-ASA was < 0.1 µg/mL. Correlation coefficients exceeded 0.99 (5-ASA, r2 = 0.994; N-acetyl-5-ASA, r2 = 0.999). The intra- and inter-assay coefficients of variation were < 8%.
In order to correct for variations in urinary concentration between the samples, the total 5-ASA concentration (5-ASA + N-acetyl-5-ASA) was expressed as a ratio to the urinary creatinine concentration. Urinary creatinine was measured using the Jaffe reaction32 (ABX Diagnostics, Montpellier, France), developed for a CobasBio Autoanalyser (Roche Diagnostics, Welwyn, UK). The intra- and inter-assay coefficients of variation were < 3% and < 7%, respectively.
All results are expressed as the median and range values unless otherwise stated. Between-group comparisons were made using the Mann–Whitney U-test, and the Friedman two-way analysis of variance was used to compare differences across the three dosage groups in the healthy volunteer study. Correlations were sought using the Spearman rank correlation coefficient, and categorical data were analysed using Pearson's chi-squared method. Logistic regression analysis was used to determine the independence of the variables.
Urinary drug excretion in healthy volunteers
Ten healthy volunteers (five males, five females; age, 22–46 years) completed the study. All remained well. Four volunteers admitted to missing a single dose and each repeated that phase of the study. All other volunteers reported full compliance and this was corroborated by pill counts.
The results of the total urinary 5-ASA to creatinine ratios on day 7 of each dosing period are shown in Figure 1. As expected, the total urinary 5-ASA to creatinine ratios increased with increasing oral dose [400 mg b.d., 1.3 (0.08–16.3); 800 mg b.d., 3.3 (1.1–15.2); 800 mg t.d.s., 4.8 (1.5–9.8); r2 = 0.02; P < 0.01]. At each dose, a modest intra-subject variation throughout the day was apparent [coefficients of variation: 400 mg b.d., 60% (26–99%); 800 mg b.d., 40% (13–82%); 800 mg t.d.s., 27% (9–46%)], although this was not statistically significant at any time point. The inter-subject variation was more marked, but fell as the oral dose increased [coefficients of variation: 400 mg b.d., 106% (92–119%); 800 mg b.d., 66% (48–78%); 800 mg t.d.s., 52% (41–59%)].
Although direct interpretation of the oral dose was not possible due to inter-subject variation, it was clear that, at oral doses of 800 mg b.d. and above, no subject provided a sample with a total 5-ASA to creatinine ratio of less than unity (Figure 2).
On stopping the drug, the total urinary 5-ASA to creatinine ratios fell, but this did not achieve statistical significance, at any dose, until the samples collected at least 36 h after the drug had been stopped were analysed. All samples collected on day 10 (60 h after the last dose) had undetectable levels of 5-ASA and N-acetyl-5-ASA (Figure 3).
Patient compliance study
Ninety-eight patients were studied. Fifty-one were male and the age ranged from 17 to 85 years. Sixty-two patients had ulcerative colitis, 26 had Crohn's disease and 10 had indeterminate colitis. Patients had been taking delayed-release mesalazine (Asacol) at 2.4 g/day (0.4–3.6 g/day) for 4.3 years (0.2–13 years). The full clinical and demographic details are shown in Table 1, and details of mesalazine therapy and rating scale results are shown in Table 2.
Table 1. Patient and disease characteristics
IBD, inflammatory bowel disease; NACC, National Association for Colitis and Crohn's disease.
* Defined as a Simple Clinical Colitis Activity Index of < 6 for ulcerative colitis/indeterminate colitis26 or a Harvey–Bradshaw Index of < 8 for Crohn's disease.27
Sex (male : female)
51 : 47
Education beyond age of 16 years (%)
Full-time employment (%)
Membership of patient group (NACC or IBD Club) (%)
Patients were invited to estimate their level of compliance in the range 0–100%, and the results are shown in Figure 4. Fifty-one patients (52%) reported full compliance with their prescribed regimen of maintenance 5-ASA therapy, whilst 42 patients (43%) reported regular non-compliance (defined as taking < 80% of the prescribed medication).
Patients reporting regular non-compliance were younger than compliant patients (P < 0.05), more likely to have been educated beyond the age of 16 years (P < 0.001), more likely to be in full-time employment (P < 0.05), more likely to be prescribed a three-times-a-day regimen (P < 0.05), more likely to be taking no other medications (P < 0.01) and more likely to perceive their medication as ineffective (P < 0.05). Compliant and non-compliant patients were, however, similar with respect to disease duration, activity and extent, the use of concomitant inflammatory bowel disease medications, family history of inflammatory bowel disease, membership of a patient organization (National Association for Colitis and Crohn's disease or Inflammatory Bowel Disease Club), quality of life (Short Inflammatory Bowel Disease Questionnaire) scores and rates of psychiatric morbidity (Hospital Anxiety and Depression Scale). Knowledge of the medication and experience of side-effects were also similar in both groups, as were the Trust in Physician Scale scores and the time since last out-patient visit. Logistic regression analysis, however, revealed that the only independent predictors of self-reported non-compliance were three-times daily dosing regimens [odds ratio (OR), 3.1; 95% confidence interval (CI), 1.8–8.4] and full-time employment (OR, 2.7; 95% CI, 1.1–6.9). Of the 63 patients who were prescribed mesalazine three times daily, 57% admitted non-compliance and, of the 47 patients in full-time employment, 62% admitted non-compliance.
All patients provided a spot sample of urine at the time of clinic attendance. The results of urinary drug measurement are shown in Figure 5. Twelve samples (12%) contained no detectable 5-ASA or N-acetyl-5-ASA. Patients providing these samples were more likely to have Crohn's disease (P < 0.01), to have poorer quality of life scores (Short Inflammatory Bowel Disease Questionnaire) (P < 0.01) and to have anxiety and/or depression as assessed by the Hospital Anxiety and Depression Scale (a score of more than eight on either sub-scale: anxiety, P < 0.01; depression, P < 0.001). Logistic regression analysis, however, showed that depression was the only independent predictor of failure to take any medication (OR, 10.5; 95% CI, 1.8–79.0). Of the 16 patients scoring more than eight on the depression sub-scale of the Hospital Anxiety and Depression Scale, 42% had undetectable urinary drug levels.
In addition to the 12 patients who had undetectable urinary drug levels, 18 patients (18%) had urinary total 5-ASA to creatinine ratios lower than expected from the healthy volunteer data (ratio of ≥ 1 in patients prescribed ≥ 1.6 g/day). Such patients were younger (P < 0.01) than those providing samples with expected levels, more likely to be in full-time employment (P < 0.05) and more likely to be prescribed three-times-daily regimens (P < 0.01). However, three-times-daily dosing was the only independent predictor of such partial non-compliance (OR, 3.7; 95% CI, 1.8–8.9). Urinary drug levels appeared to be unrelated to disease activity.
Seven of the 12 patients (58%) providing samples containing no detectable 5-ASA or N-acetyl-5-ASA admitted to some degree of non-compliance, but only two (16%) admitted to complete non-compliance. Thirteen of the 18 patients (72%) providing urine samples suggesting partial non-compliance admitted to less than 80% compliance.
Non-compliance is recognized as a key determinant of treatment failure in a variety of clinical settings.25,33–36 The causes of non-compliance, however, are often unclear, and a number of factors may contribute.14–19 Compliance data in patients with inflammatory bowel disease are limited.
In the present study, on direct enquiry, 48% of patients admitted to regularly missing doses of medication, and 43% estimated their regular compliance to be less than 80% of the prescribed amount, with 2% admitting to taking no medication at all. Three-times-daily dosing and full-time employment were identified as independent risk factors for non-compliance. Urinary drug analysis, on the other hand, revealed no detectable drug or metabolite in 12% of patients, and an additional 18% had urinary levels suggesting partial non-compliance. Complete failure to take medication was associated with depression, and partial compliance with three-times-daily dosing. Direct patient enquiry identified most of the patients who were judged to be partially non-compliant on the basis of urinary drug testing (13 of 18), but only two of 12 who were completely non-compliant, although five of these admitted to partial non-compliance.
The prevalence of non-compliance in this study is similar to that reported in other chronic disease states19, 21, 23, 33, 34, 37 and in the few previous studies of compliance in patients with ulcerative colitis.10–12 Van Hees and Van Tongeren, who also measured drug levels, reported partial compliance with maintenance sulfasalazine therapy in 41% of patients, and total non-compliance in 12%.10 Kane et al., studying pharmacy records, found that 60% of patients with quiescent ulcerative colitis collected less than 80% of prescribed medication. In this study, male sex and more than four concomitant medications appeared to increase the risk of poor compliance, whereas recent endoscopy and being married were associated with good compliance.11 In a more recent study, Nigro et al., using patient self-reporting, found that 7% of patients admitted complete non-compliance and 10.5% reported partial non-compliance. Psychiatric morbidity and disease severity were risk factors for poor compliance, whereas a longer duration of disease was associated with better compliance.12
In the present study, three-times-daily dosing was the most significant risk factor for partial non-compliance. This association is perhaps not surprising as, in other disease states, compliance is similar when medications are prescribed once or twice daily, but falls when treatments are given in three or more daily doses.18–20 The current recommendation for sulfasalazine dosing, in both active and quiescent disease, however, is still four times daily, although many use the drug twice daily. Dosing recommendations with the newer aminosalicylates vary somewhat. All recommend divided daily dosing for both active and quiescent disease, some specifying three-times-daily dosing in active disease, and some, but not all, specifying twice-daily dosing for maintenance of disease remission.38
Pharmacokinetic studies show that peak serum 5-ASA and N-acetyl-5-ASA levels are similar following once-daily or divided daily dosing with delayed-release mesalazine,39 suggesting that dose-related toxicity is likely to be similar with the two regimens. Furthermore, rectal mucosal levels of 5-ASA and N-acetyl-5-ASA, the probable determinant of therapeutic success, are similar following once-daily or divided daily dosing, suggesting that the clinical efficacy should also be equivalent.39 Once- or twice-daily dosing therefore seems to be both feasible and desirable.
The association of partial non-compliance with full-time employment was somewhat surprising, particularly given its apparent independence from the effect of three-times-daily regimens, suggesting that the inconvenience of taking medication at work was not the explanation. Whether this relates to differences in reporting, the influence of a busier lifestyle or other factors is unclear.
The association of complete non-compliance with depression is to be expected, and confirms the observations of Nigro et al., who found that psychiatric disorders, including depression and anxiety, were present in five of six completely non-compliant patients and in six of nine partially compliant patients with inflammatory bowel disease.12 Depression is not uncommon in patients with inflammatory bowel disease,40,41 and may go unrecognized. The identification of depression is clearly important as appropriate treatment may not only improve psychiatric morbidity, but may improve compliance. In the present study, the patient-rated Hospital Anxiety and Depression Scale proved to be a convenient and useful tool for the detection of often unrecognized depression.
Whilst the impact of complete non-compliance on the disease process is clear,42 the influence of partial non-compliance is less certain. An increased risk of relapse seems likely,13 although the efficacy of intermittent dosing regimens in clinical trials suggests that the impact may vary between patients, and may relate to patterns of non-compliance.43,44 The effect of partial compliance on cancer risk is similarly unknown. The imprecise definition of compliance used in studies examining the effect of mesalazine on cancer risk (the absence of documentation of non-compliance in the medical notes) suggests that even partial compliance with maintenance mesalazine therapy may be protective.7–9 Further studies in this area are clearly desirable.
It is important to acknowledge that our study is likely to have underestimated the prevalence of non-compliance. Firstly, we studied compliance at the time of clinic attendance and, although patients were unaware of the study prior to arrival, compliance is known to improve around the time of clinic attendance.45 Secondly, we studied only clinic attenders, and it seems likely that non-attenders are more likely to be non-compliant. Thirdly, we studied urinary drug levels in healthy volunteers in an attempt to give a more objective measure of compliance. Although this is clearly of value when drug levels are undetectable, the interpretation of low drug levels is difficult given the marked variability in drug excretion at a given dose. We deliberately chose a low cut-off point to define non-compliance, but appreciate that this will probably have led to an underestimation of the frequency of partial non-compliance.
In conclusion, the results of this study show that non-compliance with maintenance mesalazine therapy is widespread in patients with inflammatory bowel disease. Direct enquiry revealed partial non-compliance in 48% and urine testing revealed complete non-compliance in 12%. Three-times-daily dosing and full-time employment emerged as independent risk factors for partial non-compliance and depression as an independent risk factor for complete non-compliance. Direct enquiry detects most patients who are non-compliant, but few patients with undetectable urinary drug levels admit to complete non-compliance. These factors should be considered when choosing a maintenance drug regimen and when monitoring compliance.
We wish to express our gratitude to our colleagues in the Sheffield Teaching Hospitals NHS Trust who allowed us to study their patients, and to Swann Morton for financial support.