OUP user menu

Journal of Crohn's and Colitis: 10 (8)


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


Published on behalf of

Adalimumab dose escalation and dose de-escalation success rate and predictors in a large national cohort of Crohn's patients

Filip Baert, Elien Glorieus, Cathérine Reenaers, Geert D'Haens, Harald Peeters, Dennis Franchimont, Olivier Dewit, Philippe Caenepeel, Edouard Louis, Gert Van Assche,
DOI: http://dx.doi.org/10.1016/j.crohns.2012.03.018 154-160 First published online: 1 March 2013


Background and aims: Adalimumab is efficacious in inducing and maintaining remission in Crohn's disease but dose escalation is needed in 30–40% after 1 year. Attempts for dose de-escalation have not been studied. This study aimed to assess the need for, predictors, and outcome of dose escalation and de-escalation in a large cohort of adalimumab treated Crohn's patients.

Methods: All consecutive patients treated with open label adalimumab for active Crohn's disease from the participating centres were included in this cohort study. A detailed retrospective chart review was performed to look for possible factors predicting outcome.

Results: Eighty four percent of 720 patients had a primary response and were followed up for a median of 14 months. Thirty four percent needed escalation after a median of 7 months (0–55 months). Multivariate predictors for dose escalation were the following: prior anti-TNF use (p < 0.0001), no concomitant azathioprine or < 3 m (p < 0.02) and abnormal CRP at start (p < 0.05). Dose escalation re-induced response for at least 6 months in 67%. Only abnormal CRP at start correlated with failure of dose escalation (p = 0.02). Dose de-escalation was attempted in 54% and was successful in 63%. After a median follow-up of 14 m adalimumab was discontinued in 29% of patients.

Conclusion: In this study real life nationwide cohort of Crohn's patients treated with adalimumab dose escalation was needed in 34% and was successful in 67%. Dose de-escalation was attempted in 54% and was successful in 63%. Overall 71% of patients maintained long term response on adalimumab.

  • Adalimumab
  • Dose escalation
  • Dose de-escalation
  • Crohn's disease

1 Introduction

Adalimumab has been widely used in clinical practise for induction and maintenance therapy in moderate to severe Crohn's disease both in academic and regional hospitals since 2007. As reported in the registration trials (Classic1 and Gain2), patients responding to the initial induction regimen and subsequently loosing response to a 40 mg SC/2 weeks regimen often respond to dose escalation to weekly injections

The incidence of the need for dose escalation varies widely across cohorts from 16% to 62%.38 However, these figures are reported mainly from academic centres or referral centres. Therefore, they may overestimate the true incidence in real life clinical practise. In addition risk factors for dose escalation have not been well determined.

As the need for dose escalation is considerable and doubles the cost of treatment when continued indefinitely, many centres have tried dose de-escalation. However, data on success rates of dose de-escalation have not been reported in the literature so far.

The aim of this study was to study the incidence and success rate of dose escalation and dose de-escalation including predictors of response to dose escalation and dose de-escalation in a large real life nationwide cohort of unselected moderately to severe Crohn's patients.

2 Methods

All consecutive patients treated with open label adalimumab for active Crohn's disease according to the Belgian reimbursement criteria from the participating BIRD (Belgian IBD Research and Development, a research community comprised of all 9 National Academic centres and 17 regional hospitals with IBD focus) centres were included in this cohort study. All patients were treated according to the label of severe Crohn's disease refractory and intolerant to at least one course of steroids and/or immunomodulators. Crohn's disease patients started on adalimumab in placebo controlled trials were excluded. Only the first course of adalimumab treatment was analysed in this study. Data were collected and analysed until the end of adalimumab treatment or end of the study follow-up (September 2010). Patients were retrieved through local database searches but completeness of the cohort was confirmed through a centralised database linked to drug reimbursement procedure in Belgium.

A detailed retrospective chart review was performed to look for demographics, disease location and disease behaviour (both according to Montreal classification – worst severity ever – and at start of adalimumab), smoking, prior and concomitant medication, and CRP. In addition serious adverse events were recorded as well as the reason for adalimumab discontinuation if available. Data were captured at the following time points: at start of adalimumab; at week 4 (primary response); at the time of dose escalation; at 3 and 6 months after dose escalation; at attempt of dose de-escalation; at the end of follow-up and or of adalimumab treatment. In this retrospective study the primary outcome was clinical response as assessed by the principal investigators from the different BIRD centres using a standard clinical evaluation mainly using clinical criteria (as included in the Harvey Bradshaw index) and CRP values (in the patients with elevated CRP).

For the statistical analysis we collected CRP values at the different key treatment decision time points as a potential predictive factor retrospectively.

The following arbitrary definitions were used:

Primary response
Clinically meaningful response (i.e. clear improvement in symptoms and drop in CRP if elevated at baseline) at week 4 through week 8. Absence of primary response precluded the patient from further analysis.
Dose escalation
Patients fulfilling the criteria of primary response to 40 mg adalimumab every 2 weeks with subsequent loss of response again (i.e. reappearance of (more) symptoms and CRP re-elevation if applicable) requiring dose escalation to weekly injection as judged by the treating investigator. Only patients treated for at least 3 months were analysed. If a patient required dose escalation within the first 3 months after starting adalimumab this was considered a primary non-responder and this patient was excluded from further analysis.
Successful dose escalation
A durable response after dose escalation to adalimumab 40 mg weekly for at least 6 months or if the patient could be de-escalated successfully (see definition below) before that time period.
Successful dose de-escalation
Reduction of the dosing to 40 mg every other week for at least 6 months with durable response after prior successful dose escalation.

If a patient was dose de-escalated after less than 6 months without success this patient was not considered a failure.

The patients with insufficient follow up (< 6 months) after dose escalation or dose de-escalation were excluded from the analysis.

Response rates and need for dose escalation were compared for high volume centres (> 50 patients included) versus low volume centres (< 50 patients included in this study).

All data were entered in a central database and analysed using SPSS version 17.0. The statistical analysis compiled descriptive statistics including percentage for discrete variables, mean and median, and range for continuous variables. Kaplan Meier survival analysis was used to describe time to event analyses. Univariate and multivariate analyses were performed to look for factors predicting response, the dose escalation, and success of dose escalation and dose de-escalation. The results from the univariate analysis were expressed as odds ratios (OR) with 95% confidence intervals (95% CI). Multivariate analysis was performed using generalised linear models. Statistical significance was inferred at the 0.05 level.

The study was approved by University of Leuven Ethics committee.

The study was supported by an unrestricted grant from Abbott Belgium.

3 Results

Seven hundred and twenty consecutive patients (61% F) from 26 centres (9 academic centres) were included in this study. In this nationwide cross sectional cohort study 304 patients were included from 20 low volume centres (range 2–48 patients) and 416 patients from 6 high volume centres (range 52–102 patients). The median time of follow-up on adalimumab treatment was 14 months (range 3 to 75 months). Patient demographics and disease characteristics are depicted in Table 1. Sixty four percent of patients (n = 466) had prior exposure to anti-TNF (infliximab alone n = 415, certolizumab alone (n = 6), infliximab + certolizumab n = 45). Loss of response (46.2%) and hypersensitivity reactions (17.4%) were the main reasons for switching to adalimumab, whereas primary non-response was present in only 8.7%. Thirty eight percent were on concomitant immunomodulators (31% azathioprine/10% methotrexate) and 33% on steroids at the start of adalimumab. The 160/80 mg induction regimen was used in 73% of patients. In 12% of patients 80 mg was injected at start and thereafter 40 mg SC every other week was used. In the remainder the exact induction scheme was unknown.

View this table:
Table 1

Patient demographics.

n = 720 patients
Gender (female)61%
Median follow up (months)14(3–75)
Age at diagnosis ( median years)24(2–76)
Disease duration (median years)11(0–44)
Prior surgery for Crohn's disease41%
Disease duration in relation to last surgery (if any) (years)7(0–42)
Smoking (% yes, never, former)39, 1–46, 5–14,4
Previous anti-TNF64.7%
Infliximab alone (n = 415)57.8%
Certolizumab alone (n = 6)0.8%
Infliximab + Certolizumab (n = 45)6%
Reason for discontinuation of first anti-TNF64%
Non response9%
Loss of response46%
Infusion reaction17%
Concomitant immunomodulators at start
Concomitant steroid use at start33%
Montreal classification
A1/A2/A3 (%)a10/77/13
Disease location L1/L2/L3/L4 (%)b27/22/51/0.3
Disease behaviour B1/B2/B3 (%)c41/27/33
Disease behaviour at start ADA: B1/B2/B3 (%)d67/18/15
Perianal disease (p)42%
Concomitant SpA13%
  • a Age at diagnosis, A1: 0–16 years, A2: 16–40 years, A3: > 40 years.

  • b Maximal location of disease, L1: pure ileal disease, L2: pure colonic disease, L3: ileal and colonic disease, L4: upper gastrointestinal tract disease.

  • c Maximal disease behaviour (i.e. behaviour at worst stage of Crohn's disease history), B1: inflammatory, B2: fibrostenosing, B3: penetrating, fibrostenosing.

  • d Disease behaviour at inclusion of the study.

Fig. 1 depicts schematically the flow of the patients in the study with regard to response, dose escalation and dose de-escalation. A primary response was observed in 605/720 patients (84%). Primary response was higher 96% (215/225) in anti-TNF naïve patients than in those with prior anti-TNF treatment 85% (372/437) (p < 0.001). The primary response rate in academic centres was 86% (311/363) versus 93% (289/313) in non-academic centres but this was only significant at univariate analysis (Table 2).

Fig. 1

Flow chart of patient analysis.

View this table:
Table 2

Factors predicting primary response.

Univariant analysis:
ParameterPrimary responseOR (95% CI)Chi²Gamma (p-value)
Prior infliximab %Y 85%N 96%0.268 (0.138–0.518)1.225(< 0.001)
Type of centre (university)Y 86%N 93%0.488 (0.293–0.812)7.4650.006
No primary responsePrimary response
Behaviour maxa (B1,B2,B3)b25/15/35248/161/179− 0.234 (0.026)
Behaviour at start adalimumab (B1/B2/B3)b39/15/19348/100/74− 0.309 (0.010)
Multivariate analysis:
ParameterNp-ValueOR95% CI
No prior infliximab625< 0.0011.1071.051–1.165
  • a According to Montreal classification: maximal disease behaviour (i.e. behaviour at worst stage of Crohn's disease history).

  • b B1: inflammatory, B2: fibrostenosing, B3: penetrating, fibrostenosing.

Dose escalation was needed to maintain clinical response in 34% of patients (208/605) after a median of 7 months (range 0–55 months). According to the Kaplan Meier curve analysis, dose escalation rates were 24% over 1 year and 55% over 2 years after start of adalimumab (Fig. 2). Several factors predicted the need for dose escalation in univariate analysis including the following: perianal disease (42% versus 30% for patients without perianal disease, p = 0.003), prior anti-TNF use (44% versus 19.6%, p < 0.0001), high volume centres (40.6% for centres with > 50 patients included versus 27.2% in centres with less than 50 patients, p = 0.001), and no or less than 3 months of concomitant azathioprine at the start of the adalimumab treatment (38% versus 26% in patients on concomitant azathioprine > 3 months, p = 0.001). University centres and a high induction dose scheme was just borderline significant at univariate analysis for a higher need for dose escalation. However the only predictors for dose escalation in multivariate analysis were prior anti-TNF use (p < 0.0001), no azathioprine at start or less than 3 months (p = 0.017), and abnormal CRP at start (p = 0.025) (Table 3).

Fig. 2

Kaplan Meier survival curve of patients maintaining remission on adalimumab biweekly over time.

View this table:
Table 3

Factors predicting need for dose escalation at univariant and multivariate analysis.

Univariant analysis
Parameter% EscalationOR (95% CI)Chi²p-Value
Perianal diseaseY 42% (96/223)N 30% (100/328)1.723 (1.209–2.456)9.140.003
Prior Anti-TNF (%)Y 44% (164/365)N 19% (40/208)3.427 (2.293–5.121)38.1720
Induction dose (% 160/80)Y 39% (165/426)N 26% (18/69)0.558 (0.315–0.989)4.0750.044
High volume centreY 41% (136/335)N 27% (67/246)1.826 (1.280–2.605)11.1390.001
Centre univ vs non univY 39% (119/303)N 32% (89/282)0.713 (0.507–1.003)3.7930.051
Azathioprine < 3 monthsY 38% (162/422)N 26% (40/152)0.573 (0.380–0.864)7.1420.008
Esc noEsc yesGamma (p-value)
Location (L1/L2/L3/L4/L2 + 4/L3 + 4/L1 + 4)115/68/111/1/5/9/1135/53/106/0/2/8/40.150 (0.025)
Azathioprine/6MP at start (no/< 3 months/≥ 3 months)236/24/112155/7/40− 0.289 (0.001)
Multivariate analysis
Parameternp-ValueOR95% CI
Prior anti-TNF exposure361< 0.0010.7910.751–0.830
CRP elevation at start3610.0250.8880.843–0.932
No(or < 3 m) of azathioprine3610.0171.1391.082–1.160

Dose escalation re-induced clinical response for at least 6 months in 67% of patients (139/208). In multivariate analysis the only factor predicting (Table 4) failure of escalation was abnormal CRP at start (p = 0.020). The success rate of dose escalation was not different in patients with prior anti-TNF exposure compared to anti-TNF naïve patients or in azathioprine naïve patients versus those on azathioprine for more than 3 months (data not shown).

View this table:
Table 4

Factors predicting success of dose escalation at univariant and multivariate analysis.

Univariant analysis
Parameter%Successful escalationOR (95% CI)Chi²p-value
Curative surgery (%)Y 83% (57/69)N 68% (42/62)2.262 (0.997–5.131)3.910.048
Elevated Crp at startY 69% (64/93)N 93% (27/29)0.163 (0.036–0.734)6.880.009
Steroids and/or Immunomodulators at escalation (% no)Y 81% (93/115)N 63% (42/67)0.397 (0.202–0.784)7.3070.007
Steroids at escalation (% no)Y 77% (114/148)Y 60% (21/35)0.447 (0.206–0.973)4.2410.039
Azathioprine (> 3 months)Y 100% (5/5)N 91% (51/56)0.911 (0.839–0.989)0.4860.486
Multivariant analysis
ParameterNp-valueOR95% CI
CRP at start920.0211.2901.039–1.601

Dose de-escalation to every other week after successful escalation to weekly dosing was attempted in 75/139 (54%) patients after a median of 3 months and was successful in 63% (47/75). We found no predicting factor for success of dose de-escalation at multivariate level.

The most frequently reported serious adverse events included Crohn's exacerbations and infections (Table 5). One case of renal cell carcinoma was observed and one sudden death for unknown reason occurred during this study. No new safety signals were detected in this large cohort of patients.

View this table:
Table 5

Serious adverse events (with hospitalisation).

Gastroenteritis (n = 1)
Clostridium infection (n = 1)
Acute viral syndrome with fever (n = 1)
Herpes zoster/Zona (n = 1)
Pneumonia (n = 2)
Exacerbation of Crohn's disease:
Flare-up (n = 4)
Small bowel obstruction (n = 5)
Perforation (n = 2)
Abscess (n = 6)
Actively drainage fistulae (n = 1)
Elective surgery (n = 3)
Pulmonary embolism (n = 1)
Ulcerative skin lesions (n = 1)
Renal cell carcinoma (n = 1)
Multiple sclerosis (n = 1)
Stroke (n = 1)
Sudden death (unknown cause) (n = 1)

Overall after a median follow-up of 14 months adalimumab was discontinued in 29% of patients (n = 209). (Fig. 3). The reasons for stopping adalimumab treatment were primary non response (n = 46 22%), loss of response (n = 52 25%), adverse events (n = 50 24%), pregnancy (n = 12 6%), remission (n = 5 2%), patient's decision (n = 12 5.7%), return to infliximab (n = 6 2.9%) and missing data (n = 26 12.4%).

Fig. 3

Kaplan Meier survival curve of patients on adalimumab (any dose).

4 Discussion

In this large nationwide real life series of Crohn's patients treated with adalimumab we observed a need for dose escalation in 34% after a median follow up of 14 months. Prior anti-TNF use, abnormal CRP at start and no concomitant immunomodulator treatment (or given for less than 3 months before starting adalimumab) were predicting loss of response on the conventional bi-weekly dosing.

This nationwide cohort comprises a mix of high volume referral centres and regional centres. The high proportion of previous anti-TNF failures may account for the 34% dose escalation. This is in line with all the other series showing a good response to adalimumab in second line therapy but with a greater need for dose escalation.47 However dose escalation was equally successful in first line treated patients compared to those with prior anti-TNF use. The need for dose escalation occurs gradually over time with a median time to dose escalation of 7 months.

An interesting new finding of this study is the fact that patients without immunomodulator co-treatment needed dose escalation more frequently. Although it has been demonstrated clearly that secondary loss of response is at least partly due to immunogenicity in infliximab treated patients, data for adalimumab is still scarce. Abstracts recently presented from other series report similarly beneficial effects for immunomodulator co-treatment. One study showed a lower loss of response rate and the other a lower need for dose escalation with concomitant immunomodulator treatment.8,9 In rheumatology it has been clearly demonstrated that anti-adalimumab antibodies occur more frequently in patients without concomitant immunomodulators and they correlate with loss of response10 as was observed with infliximab in the Sonic trial.11 It should also be noted that the proportion of patients on concomitant immunomodulators in this cohort gradually diminished over time. This reflects probably the practise of stopping co-treatment upon achieving complete remission.

An abnormal CRP level only at start of the adalimumab treatment correlated with the need for dose escalation and with failure of dose escalation. However measurements of CRP at the time of dose escalation and dose de-escalation were not predictive of response to treatment adjustments.

Although a clinical response to anti-TNF treatment has usually been associated with elevated CRP,1,11,12 it was the contrary in the recently published CARE trial reporting on a prospective open label treatment with adalimumab suggesting together with our current data that a higher inflammatory burden may require higher dosage of anti-TNF.13 In addition a bias in treatment change cannot be excluded as in this retrospective cohort; in contrast to randomised controlled trials investigators were not blinded for the CRP values but may have on the contrary be guided by the CRP levels in their evaluation and treatment decisions.

Other factors such as perianal disease, high volume centres and university centres were only predictive for need for dose escalation at univariant level most likely reflecting a more refractory patient population. Somewhat contradictory is that fact that a high induction dose scheme was also borderline significant at univariate analysis for a higher need for dose escalation. We think that this result is biased. Only in a small minority was the lower induction dose used. This is likely to be the case in patients with milder Crohn's activity.

This is the first series, to our knowledge, reporting on dose de-escalation. In this retrospective study de-escalation was not done per protocol. However, it was successful in half of the patients. This is in line with dose optimisation with infliximab where in routine practise the dosing interval can often be increased again upon re-inducing response by shortening the interval. We found no factors predicting the success of de-escalation but this may be due to the small subgroups.

Almost all patients were started on a high dose induction regimen (160 mg–80 mg, 40 mg 2QW) provided by the company as negotiated by the Belgian reimbursement authorities. This may explain the high primary response rate in this cohort along with the open label real life treatment situation and with the arbitrary definition of primary response in this study. However this high primary response rate is in line with other reported open label single centre and multi centre series.47 Primary response was higher in anti-TNF naïve patients compared to patients with prior infliximab treatment reflecting a more refractory disease for the latter patients. Patients in academic centres had a lower response rate compared to patients in non academic centres. This analysis was no longer significant when corrected for primary versus secondary TNF treatment suggesting that more patients in academic centres were started on adalimumab as a second TNF treatment.

Limitations of this study include its retrospective and non-interventional nature. All patients in this study were treated in a standard fashion according to the Belgian reimbursement criteria and to the discretion of the physicians and the patients. Therefore investigator or centre bias as to when dose escalation or dose de-escalation was decided or to when concomitant immunomodulators were needed is an intrinsic risk. Therefore although no strict prospective treatment algorithms were used in this study we believe that this report closely reflects the real life situation and adds information to the randomised controlled trials conducted with adalimumab particularly on dose de-escalation.

In summary in this real life nationwide cohort of Crohn's patients treated with adalimumab dose escalation was needed in a third of patients and was successful in two thirds. Abnormal CRP at start is predictor for need of dose escalation and failure of dose escalation. However if dose escalation was successful, dose de-escalation was successful in 60%.


  • This study was sponsored through a grant from Abbott Belgium. This grant allowed for data entry and statistical analysis by the authors and their collaborators. The study was designed by independently by the authors and the BIRD and approved by Abbott international. However Abbott laboratories were not involved in the collection, analysis and interpretation of data, nor in the writing of the manuscript or in the decision to submit the manuscript for publication.

  • FB and GVA designed the study and analysed the data. EG entered all data in a database and performed the analysis. FB drafted the manuscript. GVA and EL suggested major revisions. CR, GDH, HP DF, OD, EL, Ph C were major recruiters in the study. All authors read and approved the final version of the manuscript.

  • We thank Debby Laukens and Bert Vander Cruyssen for their help with the statistical analysis.

  • Most of all we want to acknowledge all colleagues and patients from the different BIRD centres for participating in the study for their voluntary and uncompensated dedicated efforts in retrieving all the data from the charts and allowing to share these data : AZ Groeninge Kortrijk (F. D'Heygere, C. George), AZ Sint Lucas Brugge (P. Van Hootegem), AZ Sint Maarten (S. Ilegems), CHC de Liège (F. Fontaine, A. Colard), CHRVS Sambreville (N. Schoofs),CHU Liège Sart Tilman (J. Belaiche, E. Louis, C. Reenaers, C. Van Kemseke), Clinique Saint-Pierre, Ottignies (JC. Coche), Cliniques Universitaires UCL de Bruxelles (O. Dewit), Cliniques Universitaire UCL de Mont-Godinne (JF. Rahier), DDC Digestive disease centre (M. De Reuck), H Hart Ziekenhuis Roeselare-Menen (F. Baert, J. Decaestecker, D. De Wulf), Hôpital Erasme (L. Amininejad, D. Franchimont, A. Van Gossum), Imeldaziekenhuis Bonheiden (G. D'Haens), OLV Ziekenhuis Aalst (L. Du Ville, K. Hendrickx, L. Lepoutre, J. Vandervoort, P. Van der Spek), Sint-Augustinus Antwerpen (D. Sprengers, F. Van de Mierop), Sint-Jozefkliniek Bornem (P. Potvin), Universitair Kinderziekenhuis Koningin Fabiola Antwerpen (P.Bontems), UZ Antwerpen (T. Moreels, M. Van Outryve), UZ Brussel (F. Mana), UZ Gent (D. De Looze, M. De Vos, H. Peeters), UZ Leuven (M. Ferrante, P. Rutgeerts, G. Van Assche, S. Vermeire), Virga Jesse ziekenhuis Hasselt (JL. Coenegrachts), Ziekenhuis Oost-Limburg (P. Caenepeel, E. Humblet), Ziekenhuis Maas en Kempen (S. Delen), ZNA Jan Palfijn (J. Dutre), ZNA Middelheim (J. Holvoet, S. Naegels, L. Terriere).


  • 1 All Belgian centres.


View Abstract