OUP user menu

Journal of Crohn's and Colitis: 10 (10)

Editor-in-Chief

Laurence J. Egan, Ireland

Associate Editors

Shomron Ben-Horin, IsraelSilvio Danese, ItalyPeter Lakatos, HungaryMiles Parkes, UKJesús Rivera-Nieves, USABritta Siegmund, GermanyGijs van den Brink, NLSéverine Vermeire, Belgium

6.585
5.586

Published on behalf of

Appropriateness and long-term discontinuation rate of biological therapies in ulcerative colitis

Michel H. Maillard, Murielle Bortolotti, John-Paul Vader, Christian Mottet, Alain Schoepfer, Jean-Jacques Gonvers, Bernard Burnand, Florian Froehlich, Pierre Michetti, Valérie Pittet
DOI: http://dx.doi.org/10.1016/j.crohns.2013.12.026 825-834 First published online: 1 August 2014

Abstract

Background Anti-TNFα agents are commonly used for ulcerative colitis (UC) therapy in the event of non-response to conventional strategies or as colon-salvaging therapy. The objectives were to assess the appropriateness of biological therapies for UC patients and to study treatment discontinuation over time, according to appropriateness of treatment, as a measure of outcome.

Methods We selected adult ulcerative colitis patients from the Swiss IBD cohort who had been treated with anti-TNFα agents. Appropriateness of the first-line anti-TNFα treatment was assessed using detailed criteria developed during the European Panel on the Appropriateness of Therapy for UC. Treatment discontinuation as an outcome was assessed for categories of appropriateness.

Results Appropriateness of the first-line biological treatment was determined in 186 UC patients. For 64% of them, this treatment was considered appropriate. During follow-up, 37% of all patients discontinued biological treatment, 17% specifically because of failure. Time-to-failure of treatment was significantly different among patients on an appropriate biological treatment compared to those for whom the treatment was considered not appropriate (p = 0.0007). Discontinuation rate after 2 years was 26% compared to 54% between those two groups. Patients on inappropriate biological treatment were more likely to have severe disease, concomitant steroids and/or immunomodulators. They were also consistently more likely to suffer a failure of efficacy and to stop therapy during follow-up.

Conclusion Appropriateness of first-line anti-TNFα therapy results in a greater likelihood of continuing with the therapy. In situations where biological treatment is uncertain or inappropriate, physicians should consider other options instead of prescribing anti-TNFα agents.

Keywords
  • Ulcerative colitis
  • Infliximab
  • Adalimumab
  • Certolizumab pegol
  • Appropriateness
  • Treatment discontinuation

1 Introduction

Ulcerative colitis (UC) is a chronic relapsing intestinal disease associated with impaired quality of life and significant morbidity.1 A large proportion of patients can be treated with conventional therapies such as 5-aminosalicylates or immunomodulators and biological therapies are currently recommended for drug-refractory UC or acute severe flares.2 In acute severe UC, anti-TNFα therapies have been associated with a reduced colectomy rate and a similar rate of lack of efficacy compared to conventional approaches such as intravenous ciclosporin.3,4 Despite significant improvements in disease control and overall risk of severe complications that may necessitate colectomy,4 a substantial percentage of patients do not benefit from anti-TNFα treatments due to primary non-response, side effects or secondary loss of response.5 Moreover, the long-term clinical benefit of second-line anti-TNFα agents still remains elusive due to low response and remission rates.6 The colectomy rate still remains quite high in the era of biologicals, indicating a large number of unmet needs.7

Treatment guidelines for UC taking account of known levels of evidence from randomized controlled trials are useful resources to support clinicians in their daily practice and decision-making. However, results from these trials do not cover the whole spectrum of clinical scenarios that clinicians face in daily practice. To fill this gap, we developed explicit appropriateness criteria for treatment of UC using the RAND Appropriateness Method (RAM). RAM is a well-accepted method that combines available evidence based on an extensive literature review and the clinical expertise of a panel of experts.8 This method was formerly used to develop appropriateness criteria for treatment or procedures in diverse clinical situations.912 Application of criteria to data from prospective cohorts has already been carried out, for example to assess appropriateness of treatment for Crohn's disease13,14 at specific time-points. However, few attempts have to date been made to address whether appropriate treatment is actually associated with a better outcome.

The goal of the present study was to assess the appropriateness of biological therapies for UC patients in the Swiss IBD cohort Study (SIBDC). Secondary objectives were to study factors associated with appropriateness, and to analyze discontinuation of anti-TNFα treatment over time, according to appropriateness of treatment, as a measure of outcome.

2 Methods

2.1 Study design and population

This is a longitudinal study using data from the SIBDC, a national prospective clinical cohort of IBD patients enrolled since 2006.15 Patients were recruited through their gastroenterologist in private practices, regional hospitals and tertiary centers, and were followed-up annually. Inclusion criteria were a diagnosis of IBD established at least 4 months before inclusion and confirmed by radiological, endoscopic or surgical assessment, or after at least one recurrence of the disease. Case ascertainment was made based on the Lennard-Jones criteria. Patients were excluded if they suffered from another form of colitis, were not regularly followed-up, had no permanent residency in Switzerland or if they refused to sign the informed consent form. The study population consists of adult patients (> 16 years old) included in the cohort from November 2006 to March 2013, with a diagnosis of UC, and having had one biological treatment at least once in their disease course (Fig. 1).

Figure 1

Flow chart and patient sample selection.

2.2 Databases and variables

Clinical reporting forms were completed by the treating gastroenterologist or by trained study nurses at the inclusion medical visit, and during annual follow-ups. Clinical data were collected through paper forms, sent to the cohort Datacentre, entered into databases, and checked for missing data and inconsistencies in order to validate datasets.

The following variables were used in the present study: patient demographics (age, gender, age at diagnosis), clinical status (Modified Truelove and Witts Activity Index (MTWAI)), disease location, disease duration, smoking status, presence of extra-intestinal manifestations, type of anti-TNFα therapy (start and stop dates), concomitant use and history of exposure to treatments (topical 5-ASA, oral 5-ASA, oral steroids, immunomodulators) and reasons for anti-TNFα treatment discontinuation. Age, duration of disease and MTWAI score were treated as continuous variables and disease location was treating as a categorical variable, according to the Montreal classification. Therapy discontinuation was defined as the cessation of the first anti-TNFα therapy. A patient who took the same anti-TNFα for a second time or who switched to another anti-TNFα therapy was thus already censored, and not considered for these analyses. Reasons for stopping were categorized as lack of efficacy (failure due to breakthrough/loss of response or primary non-response), side effects, patient or physician wish (patient wish, conception, treatment no longer needed) and unknown.

2.3 Appropriateness criteria (EPATUC)

The RAND Appropriateness Method (RAM) was used to develop explicit appropriateness criteria to assist treatment decisions for UC. The RAM has previously been used in the European Panel on the Appropriateness of Therapy for Crohn's disease (EPACT I & II) to develop detailed criteria for assessing therapy for CD. This method combines evidence from literature and existing guidelines with experts' experience and advice, thus allowing a filling of the gap where clear evidence is missing. The RAM, which does not force consensus, comprises five major steps: 1) an extensive literature review is conducted for the proposed intervention; 2) a catalog of potential indications (i.e. clinical scenarios) for the use of therapy is prepared; 3) a first individual rating of the degree of appropriateness of each scenario is made by multidisciplinary experts. Clinical scenarios reflecting current practice are rated on a 9-point scale from 1 (extremely inappropriate) to 9 (extremely appropriate); 4) a discussion on each scenario, with possible adaptations, is performed during a panel meeting with all experts; 5) the experts rerate each scenario for a second time, during the panel session. Median and disagreement index (D) are used to categorize ratings into 3 categories: appropriate (median 7–9, without D), uncertain (median 4–6, or D) and inappropriate (median 1–3, without D).

To assess appropriateness of therapy for UC, the European Panel on the Appropriateness of Therapy for Ulcerative Colitis (EPATUC), endorsed by the European Crohn's and Colitis Organisation (ECCO), convened in Annecy, France for a 2-day meeting in June 2011. It comprised 7 gastroenterologists, 2 surgeons and 2 general practitioners from throughout Europe. The extensive literature review experts used was compiled based on, and supplementing where necessary, the ECCO UC 2011 guidelines. Seven hundred and eighteen clinical scenarios were rated, structured into 13 main clinical presentations: not refractory or refractory proctitis, mild to moderate left-sided or extensive colitis, severe colitis, steroid-dependant colitis, steroid-refractory colitis, acute pouchitis, maintenance of remission, colorectal cancer prevention and fulminant colitis (see Appendix B).

2.4 Evaluation of appropriateness in UC patients

Appropriateness of the first anti-TNF treatment was assessed on the first occasion information on such a treatment appeared in the clinical reporting forms and where the start and/or stop dates of treatment were available and consistent. The clinical situation of each patient (i.e., disease severity, location, and history of treatments) was considered and assessed at this specific time-point, and classified into the corresponding clinical situations rated by the panel. We thus assessed either indication for start of treatment (i.e., where disease was still active), or indication for treatment maintenance (i.e., where disease was inactive), depending on the situation at the time-point of observation. In situations where treatment was indicated for maintenance, the history of drug prescription was taken as such and not challenged, even if it could have been inappropriate. Each case was double-checked to ensure that the treatment was correctly assessed from the starting situation, for example in the case of a very short time-window (i.e. start and stop within 1–2 months). For the present study, we only considered cases where the treatment start and/or stop dates were available and where there was no inconsistency in the dates, i.e. for which appropriateness of treatment could be assessed. If the treatment had already ceased during the window of evaluation (e.g. without documented start and stop dates), the patient was excluded from analyses. This was particularly the case of documented past therapies in the enrolment reporting form, where no dates were asked for. Baseline characteristics of patients and reasons for stopping were first presented according to the 3 distinct categories of appropriateness, i.e. appropriate, uncertain and inappropriate. We estimated the odds of the treatment being appropriate versus not appropriate (i.e. uncertain or inappropriate). This choice was motivated by our interest in investigating factors associated with appropriate treatment, i.e. those situations where the level of evidence is the highest compared to situations where it is not or where treatment is clearly inappropriate. Finally, we estimate the proportion of patients who stopped the treatment over time according to situations where the anti-TNF treatment was considered appropriate versus not appropriate (i.e. uncertain or inappropriate). Kaplan–Meier curves were drawn for treatment cessation, whatever the cause (i.e. due to lack of efficacy, or patient's or physician's decision).

2.5 Statistical analysis

Appropriateness of treatment was assessed using algorithms developed and based on EPATUC appropriateness criteria. Descriptive cross-tables were drawn for baseline characteristics of the study population, stratified according to categories of appropriateness of biological therapies. We used cross-tabulations, with numbers and percentages, for categorical variables, and chi-square tests to test the null hypothesis of no association between categories of appropriateness and categorical variables. We calculated the mean/standard deviation or median/interquartile range of continuous variables. Univariate logistic regressions were performed for assessing variables linked to appropriateness. Kaplan–Meier curves were used to illustrate time to discontinuation of treatment according to categories of appropriateness. Log-rank tests were performed for equality of failure functions between categories. Censoring occurred at the date of treatment cessation or at the end of the follow-up. Statistical significance was assessed by p-value from likelihood ratio testing < 0.05. Analyses were performed using STATA software version 12.1 (StataCorp, College Station, TX, USA).

2.6 Ethical considerations

The study was approved by the respective ethics committees of the Swiss regions from which patients were recruited. Written informed consent was obtained from each participating patient.

3 Results

3.1 Baseline characteristics

A total of 1044 UC patients were included in the cohort by March 2013. Of those, 275 (26%) had previously been exposed to an anti-TNFα treatment. For nearly a third (N = 89) of them starting and/or stopping dates were missing or inconsistent; the final number of patients considered in the present study was thus 186 (Fig. 1). Median age of patients at the time of assessment was 28 years, ranging from 18 to 77 years (Table 1). Half of the patients were diagnosed between 11 and 28 years old and half had a disease that had lasted up to 5 years. Nearly a quarter of patients treated with anti-TNFα had extra-intestinal manifestations (N = 44), in particular arthralgia/peripheral arthritis (18%). The majority of those patients had pancolitis (N = 90; 48%) or left-sided colitis (N = 80; 43%), and their disease was mild to moderately active (median: 4, IQR: 2–8). The most frequent clinical situations, i.e. met by three quarters of all patients receiving an anti-TNFα (Table 2 ), were left-sided or extensive colitis in remission obtained by anti-TNFα (N = 66, 35%), severe colitis with oral steroids not/never given or prior successful (N = 37, 20%), steroid-dependent colitis with prior azathioprine/6-mercaptopurine (AZA/6MP) treatment failure (N = 31, 17%) and steroid-refractory colitis (N = 10; 5%). Main situations where anti-TNFα was judged inappropriate were mild-to-moderate extensive or left-sided colitis with 5-ASA treatment prior successful or failure, colitis in remission obtained by steroids or active proctitis, not refractory, with prior success with oral/topical 5-ASA (N = 11). Other inappropriate situations referred to patient receiving adalimumab (severe colitis with prior success of steroids or steroid-dependent colitis with AZA/6MP never given or given with prior success; N = 4) or certolizumab (steroid-dependent colitis with prior AZA/6MP treatment failure; N = 2).

View this table:
Table 1

Characteristics of UC patients treated with anti-TNFa, at the time appropriateness was evaluated. Values are numbers and percentages unless stated otherwise.

VariableAppropriate treatmentUncertain treatmentInappropriate treatmentAll
All patients120 (64.5)49 (26.3)17 (9.1)186
Age (years) at diagnosis*28 (21.5–37.5)27 (21–40)25 (18–35)28 (21–38)
Age (years)*36 (28–47)36 (30–49)36 (33–48)36 (28–48)
Duration (years) of disease*5 (3–10)4 (2–11)10 (2–17)5 (2–11)
Female gender48 (39.3)24 (49.0)7 (41.2)79 (42.5)
Active smoker14 (11.9)4 (8.3)1 (6.3)19 (10.6)
Disease location
Pouchitis0 (0.0)0 (0.0)1 (5.9)1 (0.5)
Proctitis10 (8.3)3 (6.1)2 (11.8)15 (8.1)
Left-sided51 (42.5)24 (49.0)5 (29.4)80 (43.0)
Extensive colitis59 (49.2)22 (44.9)9 (52.9)90 (48.4)
Disease severity*3 (1–5.5)7 (5–10)6 (4–9.5)4 (2–8)
First-line Anti-TNFα therapy
Infliximab110 (91.7)44 (89.8)11 (64.7)165 (88.7)
Adalimumab10 (8.3)5 (10.2)4 (23.5)19 (10.2)
Certolizumab pegol0 (0.0)0 (0.0)2 (11.8)2 (1.1)
Concomitant medication
Topical 5-ASA15 (12.5)10 (20.4)5 (29.4)30 (16.1)
Oral 5-ASA43 (35.8)19 (38.8)7 (41.2)69 (37.1)
Immunomodulators$31 (25.8)27 (55.1)3 (17.7)61 (32.8)
Oral steroids32 (26.7)27 (55.1)4 (23.5)63 (33.9)
History of treatment
Topical 5-ASA64 (53.3)23 (46.9)9 (52.9)96 (51.6)
Oral 5-ASA92 (76.7)38 (77.6)13 (76.5)143 (76.9)
Immunomodulators#97 (80.8)40 (81.6)8 (47.1)145 (78.0)
Oral steroids105 (87.5)45 (91.8)11 (64.7)161 (86.6)
Extra-intestinal manifestations (all)22 (18.3)17 (34.7)5 (29.4)44 (23.7)
Arthralgia/peripheral arthritis15 (12.5)14 (28.6)4 (23.5)33 (17.7)
Uveitis/iritis2 (1.7)1 (2.0)0 (0.0)3 (1.6)
Pyoderma gangrenosum0 (0.0)1 (2.0)0 (0.0)1 (0.5)
Erythema nodosum1 (0.8)0 (0.0)0 (0.0)1 (0.5)
Aphtous oral ulcer/stomatitis2 (1.7)2 (4.1)0 (0.0)4 (2.1)
Ankylosing spondilitis/sacroilitis3 (2.5)1 (2.0)4 (23.5)8 (4.3)
Primary sclerosing cholangitis1 (0.8)0 (0.0)0 (0.0)1 (0.5)
  • * Median (IQR); $ azathioprine/6-mercaptopurine; # azathioprine/6-mercaptopurine/methotrexate.

View this table:
Table 2

Summary of clinical situations of UC patients that received anti-TNFα (N = 186). These situations reflect main categories and details, as considered by the EPATUC experts' panel (Appendix B).

Clinical situationN
Colitis$ in remission obtained by anti-TNFα agents, with azathioprine/6MP treatment not/never given or prior success (or failure)*33 (33)
Steroid-dependent colitis$, with prior azathioprine/6MP treatment failure31
Severe colitis$, with prior success with oral steroids23
Severe colitis$, with oral steroids not/never given14
Steroid-refractory colitis$, with azathioprine/6MP treatment not/never given or prior success (or failure)4 (6)
Steroid-dependent colitis$, with prior success with azathioprine/6MP treatment9
Proctitis in remission obtained by anti-TNFα agents, with azathioprine/6MP treatment not/never given or prior success (or failure)3 (4)
Severe colitis$, with prior oral steroids failure6
Active proctitis, refractory, with azathioprine/6MP treatment not/never given or prior success (or failure)3 (3)
Mild-to-moderate extensive colitis, with 5ASA treatment not/never given or prior success (or failure)3 (2)
Colitis$ in remission obtained by steroids, with 5-ASA treatment not/never given, prior success or prior failure2
Mild-to-moderate left-sided colitis, with oral/topical 5ASA treatment not/never given or prior success2
Active proctitis, not refractory, with prior success with oral/topical 5ASA2
Steroid-dependent colitis$, with azathioprine/6MP not/never given1
Mild-to-moderate left-sided colitis, intolerance to topical therapies1
Active pouchitis, with prior successful imidazole treatment1
  • * Definition of the drug status: Not/never given = never given or was given at a suboptimal dose, Failure = appropriate treatment in terms of doses and duration which led to an insufficient response or adverse events (currently or previously), Prior success = treatment with this drug led to remission but was stopped for various reasons; $ refers to left-sided or extensive colitis.

3.2 Appropriateness of anti-TNFα agents

Overall, anti-TNFα treatment was considered appropriate, uncertain or inappropriate in 65%, 26% and 9% of cases, respectively. These proportions did not differ for patients enrolled in tertiary centers or not (p = 0.210). Infliximab was the most frequently-prescribed first line anti-TNFα treatment (N = 165; 89%). A third of the patients treated by infliximab or adalimumab were on concomitant immunomodulators (N = 55, resp. N = 6); oral 5-ASAs were combined with an anti-TNFα agent in 37% and 42% of the cases for infliximab and adalimumab respectively. Concomitant immunomodulators (55%), oral steroids or oral 5-ASA (41%) were more frequently prescribed when the anti-TNFα treatment was of uncertain appropriateness. Among the 186 patients on anti-TNFα treatment, 68 (37%) discontinued treatment during the follow-up period (Table 3). This observation was more frequent, in proportion, for patients whose indication for biological treatment was inappropriate (p = 0.06) or uncertain (p = 0.04) compared to a biological treatment considered appropriate. The highest proportion of treatment discontinuation due to lack efficacy was observed when the treatment was considered uncertain (29% of all patients under uncertain biological therapy). Patient or physician wish to stop accounted for 36% of all cessation reasons, and was more frequent in the group of patients on inappropriate biological treatment, compared to not inappropriate (p = 0.033). The proportion of patients who were appropriately treated was significantly lower among those who experienced a lack of efficacy (Fig. 2 ), compared to those who did not (43% vs. 70%; p = 0.027). For nearly 60% of patients who stopped the first line biological treatment because of lack of efficacy, the treatment was considered uncertain or inappropriate, compared to 38% for patients who stopped due to side effects or own/physician wish.

Figure 2

Comparison of the proportions of treatment considered appropriate, uncertain or inappropriate between groups of patients who needed to stop treatment for a specific reason and those who did either stop or not stop but not for this specific reason.

View this table:
Table 3

Reasons for stopping anti-TNFα treatment (N = 68).

Reason for discontinuationAppropriate treatmentUncertain treatmentInappropriate treatmentAll
All patients1204917186
All discontinuations36 (30.0)23 (46.9)9 (52.9)68 (36.6)
Specific reason
Lack of efficacy14 (11.7)14 (28.6)4 (23.5)32 (17.2)
Side effects15 (12.5)8 (16.3)1 (5.9)24 (12.9)
Patient or physician wish15 (12.5)4 (8.2)5 (29.4)24 (12.9)

3.3 Factors associated with appropriateness of treatment

Having an appropriate anti-TNFα treatment was associated with less severe disease (OR = 0.74; p < 0.001), less concomitant use of steroids (OR = 0.38; p = 0.003) or of immunomodulators (OR; p = 0.010) (Table 4). The presence of extra-intestinal manifestations increased the odds of having a treatment which was not appropriate by 59% (p = 0.011). The need to stop treatment (whatever the cause) was lower in patients treated appropriately versus not appropriately (OR = 0.47; p = 0.018), as was the need to stop because of lack of efficacy (OR = 0.36; p = 0.011). Stop of treatment due to side effects, or because of patient or physician wish was not associated with appropriateness of treatment.

View this table:
Table 4

Crude odds ratio (OR) and 95% confidence intervals (CI) for appropriateness of anti-TNFα therapy for UC patients according to clinical and therapeutic factors.

VariableOR (95% CI) for appropriate vs. not appropriate treatmentp-value
Female gender0.78 (0.43–1.44)0.432
Age1.00 (0.98–1.02)0.962
Age at diagnosis1.00 (0.98–1.02)0.870
Duration of disease0.99 (0.95–1.04)0.826
Active smoker1.66 (0.57–4.84)0.351
Disease severity0.74 (0.67–0.82)< 0.001
Left-sided vs. extensive disease location0.87 (0.47–1.64)0.679
Proctitis vs. extensive disease location1.05 (0.33–3.35)0.933
Concomitant immunomodulators0.43 (0.23–0.82)0.010
Concomitant topical 5-ASA0.50 (0.22–1.10)0.085
Concomitant oral 5-ASA0.89 (0.48–1.65)0.717
Concomitant oral steroids0.38 (0.20–0.72)0.003
Extra-intestinal manifestations0.41 (0.20–0.82)0.011
History of oral 5-ASA1.07 (0.53–2.17)0.853
History of topical 5-ASA1.16 (0.64–2.11)0.629
History of immunomodulators1.74 (0.86–3.53)0.121
History of oral steroids1.22 (0.51–2.88)0.656
Stop of treatment (whatever the cause)0.47 (0.25–0.88)0.018
Lack of efficacy0.36 (0.17–0.79)0.011
Side effects of treatment0.93 (0.38–2.25)0.872
Patient or physician wish to stop0.93 (0.38–2.25)0.872

3.4 Appropriateness and time-to-failure of the first-line biological treatment

Cumulative first-line anti-TNFα treatment discontinuation rates differed significantly between groups of patients appropriately and not appropriately treated, (Fig. 3A). The discontinuation rate was 26% (95% CI: 18%–36%) after 2 years [resp. 47% (95% CI: 35%–61%) after 4 years] if the treatment was appropriate, compared to 54% (95% CI: 40%–68%) if not appropriate [resp. 66% (95% CI: 52%–81%) after 4 years]. In case of discontinuation due to lack of efficacy, discontinuation rates were lower after 4 years for patients treated appropriately (20%, 95% CI: 12%–32%) than after 2 years for patients not treated appropriately (33%; 95% CI: 20%–50%), (Fig. 3B).

Figure 3

Kaplan–Meier time-to-failure estimates for patients with the first-line anti-TNFα treatment. (A): estimation of the proportion of patients who stopped for any cause, (B): estimation of the proportion of patients who stopped due to lack of efficacy.

4 Discussion

We report for the first time results from the application of appropriateness criteria (EPATUC) to ulcerative colitis patients included in a cohort study. Using these criteria, we assessed the appropriateness of the first-line biological treatment in 186 UC patients from the SIBDC. We found that, for 64% of them, the use of biological treatment was considered appropriate. During follow-up, 37% of all patients discontinued biological treatment, 17% specifically because of lack of efficacy. Estimation of time-to-failure of treatment showed that patients on an appropriate biological treatment tended to stay longer on the treatment than those for whom the treatment was considered not appropriate. Furthermore, these results contribute, indirectly, to a further validation of the RAM approach.

Biological therapies have revolutionized current management strategies for UC, especially in situations where surgery was previously the only available option.4 Results from existing studies including randomized controlled trials have helped clinicians to stratify clinical situations and led to the development of treatment guidelines that aimed to standardize clinical practice.1618 Unfortunately, scientific evidence is often lacking for many clinical situations that are faced by clinicians in their daily practice. To fill this gap, appropriateness criteria for several disease presentations have been established for ulcerative colitis (Appendix B), as previously done for Crohn's disease.19 These EPATUC criteria could successfully be applied to data from patients of the SIBDC.

We found that about a quarter of UC patients followed up in the SIBDC have been exposed to biologicals. This relatively high proportion may have at least two plausible explanations. First, a large number of SIBDC patients have been followed up for more than 5 years, thus increasing the likelihood of being exposed to anti-TNFα agents. Second, since our cohort started enrolment in the era of biologicals, this might reflect a true tendency in the current management of UC. Third, a majority of patients from the SIBDC were enrolled in tertiary centers. Follow-up was, however, mostly accomplished in private practice, thus possibly limiting any inherent selection bias in this regard.

Our data show that the majority of UC patients receiving biologicals were treated appropriately. Data addressing whether early introduction of biologicals might benefit subsets of UC patients are currently lacking. Nevertheless, clinicians often face situations where severity or refractoriness of disease might lead to an early-aggressive type of approach. Such a choice might condition future decisions on treatment options and affect physicians' willingness to “step-down” to conventional therapies. Interestingly, in this study, patients had average disease duration of 4 years prior to exposure to biologicals when appropriately treated. In contrast, inappropriate treatment was associated with longer disease duration but a lower likelihood of being previously exposed to immunomodulators or steroids. Those observations suggest that Swiss clinicians tend to prescribe anti-TNFα agents prior to conventional therapies when disease duration is longer. This “top-down-like” approach was associated with a higher likelihood of discontinuing treatment during follow-up. Future studies should evaluate efficacy of a top-down strategy and discontinuation rates earlier in the course of disease to assess whether a similar trend to that found in Crohn's disease can be observed.20

Our data may also suggest that Swiss physicians often opt for biologicals in unclear indications, especially for severe and refractory cases. One could speculate that this might be related to differences between patients' and physicians' perception of post-colectomy health status and related quality of life. In this regard, surveys conducted on doctors' and patients' perception of the pre- and post-colectomy state when living with moderate colitis differed significantly.21 Physicians and post-colectomy patients estimated utility of life with moderate colitis more poorly than the post-colectomy state. In contrast, pre-colectomy patients had similar views on both health states. Hence, the observations presented here might reflect the influence of patients' willingness to continue with medical treatments on physicians' decision. Such situations might however be more adequately treated surgically. Accordingly, the risk of having an uncertain indication for biologicals was higher among patients where there was a lack of efficacy. Disease severity and concomitant use of steroids or of immunomodulators were more likely to be encountered in uncertain indications. Overall, our data might reflect a certain degree of anti-TNFα drug overuse, especially in complicated cases. Unfortunately, this practice does not seem to provide any additional benefit in terms of drug efficacy and discontinuation rates.

The percentage of patients having stopped treatment in our study compares relatively well with the numbers that have been reported in prior studies, despite a different study design. In the ACT1/2 extension studies, 30.6% patients discontinued infliximab infusions over a follow-up period of up to 3 years.22 In a small prospective study comparing long-term outcomes of infliximab- or adalimumab-treated patients, close to a quarter of the patients discontinued treatment during a follow-up period of 2 years.23 When looking at moderate-to-severe UC, infliximab therapy led to a 68% sustained clinical response rate and 17% colectomy rate after a follow-up of 33 months.5

When to stop biological treatment is a topic of much concern.24 In this study, patient's or doctor's wish accounted for slightly more than a third of all treatment discontinuations. Interestingly, the wish to stop was more frequent when treatment was considered inappropriate. In an observational Danish study, treatment discontinuation for patient's or doctor's wish while in clinical remission was observed in 24% and 30% of CD and UC patients, respectively.25 During follow-up, most patients relapsed and had to come back on treatment with a subsequent good response rate. Thus, when treatment is appropriate, i.e. highly needed, patient's and doctor's adherence is also higher. This is in line with studies showing that the perception of a high need for therapy plays a role in adherence.26

The main strengths of the present study were the ability to determine appropriateness of use of biologicals and to assess its impact on long-term discontinuation. This was particularly true due to the high number of patients under biologicals in the Swiss IBD cohort and data available to characterize them over time. The usefulness of prospective cohort data to look for significant outcomes according to appropriateness of treatment for complicated chronic diseases such as IBD is therefore once again underscored.

Our study nevertheless has several limitations. Some are related to the development of appropriateness criteria, for which certain factors were not taken into account, i.e. the presence of extra-intestinal manifestations. This might be explained by the fact that, at the time of the RAND panel, the situations where biologicals might be used and the increased attempt to use them in practice, were underestimated. Other limitations might be linked to our study sample, which only comprised those cases for which all information was available. We cannot guarantee that selection bias may not have happened. However, it might have been more probable if we had excluded groups of patients from the same practice, or sets of data collected by the same physician or study nurses. This was not the case in our study. We excluded 89 cases who had taken biologicals, but who had already stopped them at the time of enrolment in the cohort. Unfortunately, for many of them (86/89), dates were missing simply because we decided deliberately not to collect dates of past treatment at enrolment for data collection efficacy reasons. To have these data would require looking again at the charts of the patients concerned, which was not resource-feasible for the present study. Finally, the total number of patients included in this analysis was relatively limited, especially for subgroup analyses, despite the large number of UC patients in the SIBDCS.

In conclusion, our observations indicate that treatment appropriateness has clinical relevance in the long-term and that physicians should consider integrating appropriateness of care in their daily clinical practice. In the specific setting of biological therapies for UC, our data suggest that current practice should probably avoid prescribing anti-TNFα agents in situations where the indication is uncertain. They do however emphasize that clinicians facing those situations might judiciously consider switching to drugs targeting other pathways such as vedolizumab27 or opt for a surgical approach.

Competing Interest

No author has any conflict of interest or financial ties relevant to the manuscript to disclose.

Contributorship

Conception and design (1), analysis and interpretation of data (2), drafting of the manuscript (3), critical revision of the manuscript for significant intellectual content (4), final approval of the version to be published (5), obtained funding (6).

Michel H. Maillard: 1, 2, 3, 4, 5; Murielle Bortolotti: 1, 2, 4, 5; John-Paul Vader: 1, 2, 4, 5; Christian Mottet: 1, 4, 5; Alain Schoepfer: 1, 4, 5; Jean-Jacques Gonvers: 1, 4, 5; Bernard Burnand: 1, 2, 4, 5; Florian Froehlich: 1, 2, 4, 5; Pierre Michetti: 1, 2, 4, 5; Valérie Pittet: 1, 2, 3, 4, 5, 6.

Appendix A Members of the SIBDCS study group

Peter Bauerfeind; Christoph Beglinger; Stefan Begré; Dominique Belli; José Bengoa; Luc Biedermann; Janek Binek; Jan Borovicka; Christian Braegger; Patrick Bühr; Bernard Burnand; Emanuel Burri; Rafael Camara; Dominique Criblez; Philippe de Saussure; Lukas Degen; Mohamed Ali Delhavi; Joakim Delarive; Christopher Dörig; Gian Dorta; Tobias Ehmann; Ali El Wafa; Christian Felley; Markus Fliegner; Alain Frei; Pascal Frei; Remus Frei; Michael Fried; Michael Friedt; Florian Froehlich; Heiko Frühauf; Raoul Furlano; Suzanne Gallot-Lavallée; Martin Geyer; Marc Girardin; Sébastien Godat; Jean-Jacques Gonvers; Ignaz Good; Oliver Götze; Beat Gysi; Antoine Hadengue; Marcel Halama; Serge Hediger; Pius Heer; Beat Helbling; Mona Hellstern; Peter Hengstler; Denise Herzog; Cyrill Hess; Klaas Heyland; Thomas Hinterleitner; Philippe Hiroz; Claudia Hirschi; Petr Hruz; Patrick Janiak; Christian Jost; Pascal Juillerat; Vera Kessler Brondolo; Christoph Knoblauch; Claudia Krieger; Gerd A. Kullak-Ublick; Markus Landolt; En-Ling Leung-Ki; Philippe Maerten; Michel Maillard; Christine Manser; Michael Manz; Urs Marbet; Igor Marsteller; Rémy Meier; Christa Meyenberger; Pierre Michetti; David Mondada; Darius Moradpour; Patrick Mosler; Christian Mottet; Christoph Müller; Pascal Müller; Beat Müllhaupt; Andreas Nagy; Cristina Nichita; Thierry Nicolet; Hansjürg Nüesch; Andreas Nydegger; Isabelle Pache; Franziska Piccoli; Julia Pilz; Valérie Pittet; Bruno Raffa; Ronald Rentsch; Jean-Pierre Rey; Daniela Rogler; Gerhard Rogler; Markus Sagmeister; Bernhard Sauter; Mikael Sawatzki; Gerold Schacher; Michael Scharl; Nora Schaub; Martin Schellin; Susanne Schibli; Hugo Schlauri; Jean-François Schnegg; Alain Schoepfer; Franck Seibold; Mariam Seirafi; David Semela; Marc Sidler; Johannes Spalinger; Holger Spangenberger; Philippe Stadler; Peter Staub; Michael Steuerwald; Alex Straumann; Michael Sulz; Michela Schäppi; Joël Thorens; Radu Tutuian; John-Paul Vader; Stephan Vavricka; Dominique Velin; Francesco Viani; Roland Von Känel; Alain Vonlaufen; Dominique Vouillamoz; Gert Wachter; Jürg Wermuth; Paul Wiesel; Samuel Zamora; Jonas Zeitz

Appendix B Supplementary data

Supplementary material.

Appendix B Supplementary data

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.crohns.2013.12.026.

Acknowledgments

This study was carried out by all the gastroenterologists and investigators listed in the Swiss IBD Cohort Study Group. The study is supported by the Swiss National Science Foundation (SNSF) grant nos 33CS30-134274 (Swiss IBD cohort study) and 32473B-138498 (appropriateness of care in IBD).

Footnotes

  • 1 See Appendix A.

References

View Abstract