Early clinical remission and normalisation of CRP are the strongest predictors of efficacy, mucosal healing and dose escalation during the first year of adalimumab therapy in Crohn’s disease

Authors


Dr. P. L. Lakatos, 1st Department of Internal Medicine, Semmelweis University, Koranyi Sandor utca. 2/A, H-1083 Hungary.
E-mail: kislakpet@bel1.sote.hu

Abstract

Aliment Pharmacol Ther 2011; 34: 911–922

Summary

Background  Adalimumab is a fully human monoclonal antibody targeting tumour necrosis factor with proven efficacy in the treatment of Crohn’s disease (CD).

Aim  To investigate the predictors of medium-term clinical efficacy and mucosal healing during adalimumab therapy, in patients with CD, in specialised centres approved for biological therapy in Hungary.

Methods  Data capture of the 201 CD patients was standardised and prospective (male/female: 112/89, median age: 33.0 years, duration: 8 years). Previous infliximab therapy had been administered in 48% of patients, concomitant steroids in 41%, azathioprine in 69% and combined therapy in 27% of patients.

Results  Overall clinical response and remission rates at 24 weeks were 78% and 52%, respectively; at 52 weeks were 69% and 44%, respectively. Endoscopic improvement and healing were achieved in 43% and 24% of patients. In a logistic regression model, clinical efficacy and CRP at week 12, need for combined immunosuppression at induction, shorter disease duration and smoking were identified as independent predictors for 12-month clinical outcome, whereas CRP at week 12, clinical remission at week 24, inflammatory parameters and nonsmoking were associated to endoscopic improvement/healing. Intensification to weekly dosing was needed in 16% of patients. Parallel azathioprine therapy and clinical remission at week 12 were inversely associated with dose escalation.

Conclusions  Clinical efficacy and normalised CRP at week 12 (early deep clinical remission) are associated with medium-term clinical efficacy and mucosal healing during adalimumab therapy, whereas need for combined immunosuppression at induction and smoking status are predictors for non-response. Parallel azathioprine therapy may decrease the probability for dose escalation.

Introduction

Crohn’s disease (CD) is a progressive condition, with most patients developing a penetrating or stricturing phenotype over time.1 The introduction of anti-tumour necrosis factor (anti-TNF) therapies over the past 10 years, along with accumulating evidence from landmark trials and clinical practice, has led to a significant change in patient management and treatment algorithms. Over the past 10 years, treatment goals changed, allowing patients to achieve not only symptomatic relief but also sustained steroid-free remission. The anti-TNFs were demonstrated to be effective for the treatment of both luminal and fistulising disease. Scheduled therapy with both infliximab2, 3 and later adalimumab4–6 was shown to be associated with an increased likelihood of extended remission, mucosal healing in a significant proportion of patients, reduced number of hospitalisations and lower corticosteroid requirements. Long-term clinical benefit is demonstrated for both agents.7, 8 Adalimumab (Humira, D2E7; Abbott Laboratories, Abbott Park, Chicago, IL, USA), in particular, was found to be effective in refractory patients either naïve to or previously treated with, infliximab.5, 9–11

Unfortunately, approximately 25–40% of patients who respond initially to this therapy require multiple dose and interval adjustments to maintain clinical response, and the annual discontinuation rate for adalimumab therapy is approximately 10%. Therapy discontinuation is generally due to either loss of response or side effects8, 12. Nevertheless, relatively little is known about which factors are clinically relevant in predicting continued response or failure during adalimumab therapy.

In the pivotal clinical trials, it has been demonstrated that anti-TNF agents are more effective in early disease,6, 12 but investigations regarding the importance of concomitant immunosuppressive therapy or biomarkers (e.g. CRP at inclusion) have led to contradictory results.

Most data are available regarding predictors of response for short- and long-term clinical benefit during infliximab therapy. Clinical response was reported to be associated with baseline CRP for certolizumab.13 Similarly, in a recent study by the Leuven group, CD patients who had elevated baseline levels of CRP (> 3 mg/L) responded to infliximab therapy better than patients with normal levels. Survival analysis demonstrated that patients in whom CRP levels normalised early, after 4 of 10 weeks, had better long-term outcomes. The CRP level at time of endoscopy was significantly correlated with the degree of mucosal healing.14 Less data are available on the predictors for the efficacy of adalimumab therapy. The best evidence has been provided by a study conducted in Leuven,15 where besides drug-trough level, CRP kinetics were identified as a predictor for sustained clinical benefit during a median 20-month follow-up period of adalimumab therapy in CD patients who failed to respond to infliximab. Patients who had normalised CRP levels (< 3 mg/L) at both week 4 and week 12 discontinued adalimumab less frequently and showed longer sustained clinical benefit. Moreover, time to dose escalation was longer in patients who were treated with immunomodulators.

In addition, combination therapy with immunosuppressants and infliximab has shown increased efficacy in comparison with biological-only therapy in both azathioprine (AZA) naïve and AZA-exposed CD patients.16, 17 In contrast, the importance of concomitant immunosuppression was not proven for adalimumab in post hoc analysis of the randomised controlled trials.6 Moreover, previous anti-TNF exposure was associated with numerically lower clinical remission rates, although a significant difference was not achieved. Other important clinical variables that have been associated with progressive CD, and may influence the outcome of biological therapy, include young age at onset, need for repeated courses of corticosteroids, stricturing or penetrating disease behaviour, severe lesions as confirmed by endoscopy and smoking.1 In addition, smoking was inversely associated with response to short-term infliximab therapy,18 but results from other studies contradict this.

Identifying predictors of response to adalimumab in a complex model, which includes both clinical, laboratory, and treatment variables would be of great benefit in the patient selection and long-term management. Furthermore, identifying modifiable factors associated with a prolonged response could allow for optimisation of response rates and duration. Therefore, in the present study, our aim was to investigate the predictors of medium-term clinical efficacy and mucosal healing during adalimumab therapy, in patients with CD, in specialised centres approved for biological therapy in Hungary.

Materials and methods

Patients

Two hundred and one well-characterised, unrelated, consecutive CD patients (male-to-female ratio: 112/89, median age at presentation: 24 years (IQR: 19–31 years), duration: 8 years (IQR: 4–12 years)) with a complete clinical follow-up were included from specialised centres providing biological therapy in Hungary. The clinical characteristics are presented in Table 1.

Table 1.   Clinical characteristic of patients with Crohn’s disease
 CD (n = 201)
  1. * Median (IQR).

  2. † n (%), data for smoking habits are missing for three patients.

  3. ‡ High sensitivity-CRP > 10 mg/L, data are missing for five patients.

  4. § Including induction only therapy episodic therapy, or change due to patient preference.

Male/female112/89
Age at presentation (years)*24 (19–31)
Duration (years)*8 (4–12)
Familial IBD†16 (8.0%)
Location (n)
 L1 (ileal)15 (7.5%)
 L2 (colonic)72 (35.8%)
 L3 (ileocolonic)110 (54.7%)
 L4 (only upper GI)4 (3.8%)
 All L418 (8.9%)
Behaviour (n) at diagnosis
 B1 (nonstricturing nonpenetrating)85 (42.4%)
 B2 (stricturing)41 (20.4%)
 B3 (penetrating)75 (37.2%)
Perianal disease†97 (48.3%)
Frequent relapse†152 (75.6%)
Arthritis†90 (44.8%)
Liver†3 (1.5%)
Ocular†16 (8.0%)
Cutaneous†32 (15.9%)
Steroid use ever/during at the start of adalimumab†186 (92.5%)/ 83 (41.3%)
Azathioprine use ever/during at the start of adalimumab†185 (92.0%)/ 139 (69.2%)
Combined immunosuppression at the start of adalimumab (steroid and azathioprine)†53 (26.4%)
Previous anti-infliximab†97 (48.3%)
 Primary nonresponse12
 Secondary loss of response39
 Allergic reaction31
 Others§15
CRP elevated at the start of adalimumab‡131 (66.8%)
Previous resective surgery†77 (38.3%)
Smoking during adalimumab induction†42 (21.2%)

Methods

Diagnosis was based on the Lennard-Jones criteria.19 The disease phenotype (age at onset, duration, location and behaviour) was determined according to the Montreal Classification.20 Medical records, including data regarding the presence of major extraintestinal manifestations (EIM), previous frequency of flare-ups (frequent flare-up: > 1 clinical relapse /year,21) previous surgical procedures (resections or perianal procedures), the presence of familial IBD, smoking habits and perianal involvement, were determined by a thorough review of the patients’ medical charts, which had been collected in a uniform format. Previous and concomitant medical therapy (steroid and/or immunosuppressive, or previous biological therapy) was meticulously registered. Clinical data, laboratory data and endoscopy results of CD patients in whom adalimumab therapy was started during the first year of the study period was captured prospectively between 1 December 2008 and 31 December 2010. There was a change in the regulations regarding the use of biologicals in Hungary after 1 December 2008. Adalimumab became available in an out-patient setting for the treatment of CD patients with luminal (and additional fistulising) disease with a CDAI > 300 despite adequate previous conventional therapy (including appropriate use of steroids and immunosuppressives for at least 3 months; or failure to respond to, or intolerance to, either steroid or immunosuppressive therapy). Due to health authority regulations, patient follow-up appointments are mandatory at least every 3 months (regular or extraordinary visits). These visits include clinical assessment, review of patient diaries, CDAI22 (and PDAI) determination, laboratory assessment (including CRP), chest X-ray in every 6 months and in fistulising patients, objective determination of response by MRI, EUA or rectal ultrasound at 12 months. Consequently, patient selection and follow-up were standardised and uniform in all specialised centres from this date. Of note, all centres were monitored for quality of care and regulation compliance by the Hungarian Insurance Bureau (OEP) in June 2011.

Clinical response was defined as a ΔCDAI > 70 points (and a > 50 decrease of fistula drainage in patients with concomitant fistulising disease), whereas clinical remission was defined as a CDAI < 150 points (and complete cessation of fistula drainage in patients with concomitant fistulising disease). Medium-term efficacy was defined as the outcome of 1 year of therapy. Endoscopic mucosal status was assessed by the presence or absence of ulcers and erosions. Complete mucosal healing was defined as the absence of any mucosal lesions or signs of active inflammation. A marked improvement of mucosal conditions without complete healing was defined as partial healing.23 The definition of smoking consisted of smoking ≥ 7 cigarettes per week for at least 6 months.24

The central coordination and database management was completed at the 1st Department of Internal Medicine, Semmelweis University (by LSK, PLL). The study was approved by the Semmelweis University Regional and Institutional Committee of Science and Research Ethics.

Statistical methods

Variables were tested for normality using Shapiro–Wilks W test. The chi-squared test and chi-squared test with Yates correction and logistic regression analysis were used to assess the association between categorical clinical variables and clinical/endoscopic outcome. Sensitivities, specificities, positive predictive values (PPV) and negative predictive values (NPV) were calculated to determine the predictive power of hs-CRP and clinical parameters in distinguishing between CD patients reaching or not reaching clinical and endoscopic remission or improvement during maintenance therapy. Kaplan–Meier survival curves were plotted for analysis with LogRank and Breslow tests. In addition, Cox-regression analysis was used to assess the association between categorical clinical variables and probability of dose escalation. Variables with P < 0.1 were included in the multivariate testing. A P-value < 0.05 was considered significant. For statistical analysis, SPSS15.0 (SPSS Inc, Chicago, IL, USA) was used.

Results

Indication for adalimumab therapy and concomitant medication

Indication for biological therapy was active luminal disease in 63.7% of patients and active luminal and parallel fistulising disease in 36.3% of patients. Ninety-seven (48.3%) patients had been treated previously with infliximab therapy. The induction dose was 80/40 mg in 61.7% of cases and 160/80 mg in the remaining 38.3%. Concomitant immunosuppression at induction therapy consisted of steroids in 41.3% of patients, AZA in 69.2% and combined therapy in 26.4%. The hs-CRP at induction was elevated in 66.8% of patients (median hs-CRP: 17.5 (IQR: 6.7–40) mg/L).

Clinical efficacy and predictors of clinical efficacy at 24 and 52 weeks

Overall clinical response and remission rates are shown in Table 2. Need for concomitant steroids (P = 0.053), combined immunosuppression (steroids and AZA, P = 0.013) during induction therapy, previous surgery (P = 0.004), low CRP at week 12 (< 10 mg/L, P < 0.001, n = 189, patient with a complication before week 12 or a missing week 12 CRP value were excluded, median hs-CRP: 4.6 (IQR: 1.6–19.1) mg/L), and clinical efficacy (response or remission) at week 4 and at week 12 (P = 0.006 and P < 0.001) were associated with clinical response or remission at week 24 (Table 3). Clinical response at week 12 (P < 0.001), low CRP at week 12 (P < 0.001, OR: 5.03, 95% CI: 2.63–9.64) and previous surgical resection (P = 0.012, OR: 0.48, 95% CI: 0.27–0.85) were associated with the probability of clinical remission at week 24, with further two variables close to being statistically significant (need for combined concomitant immunosuppression at induction (P = 0.06, OR: 0.55, 95% CI: 0.29–1.03) and higher induction dose of adalimumab (160/80 mg, P = 0.065, OR: 1.72, 95% CI: 0.97–3.06).

Table 2.   Clinical efficacy of adalimumab therapy
 Week 4* (%)Week 12* (%)Week 24* (%)Week 52* (%)
  1. Response: ΔCDAI > 70 points from baseline.

  2. Remission: CDAI < 150 points.

  3. * n = 201 at week 4, n = 200 at week 12 and week 24 and n = 196 at week 52.

No response/stopped612.52230.6
Response63.7362625
Remission30.451.55244.4
Remission off steroids4438.8
Table 3.   Predictors of clinical efficacy of adalimumab therapy at week 24 and week 52
 Clinical efficacy at week 24P value
No response (%)Response (%)Remission (%)
Steroids at induction
 Yes30.527.145.10.053
 No16.124.456.8
Combined immunosuppression at induction
 Yes16.927.755.40.013
 No36.521.242.3
Previous surgery
 Yes23.617.958.50.004
 No19.539.041.6
Efficacy at week 4
 None5018.9230.006
 Response16.733.113.1
 Remission33.34863.9
Efficacy at week 12
 None10016.76.8< 0.001
 Response061.17.8
 Remission022.285.4
CRP < 10mg/L at week 12 (n = 189)
 Yes10.722.367< 0.001
 No36.835.327.9
 Clinical efficacy at week 52P value
No response (%)Response (%)Remission (%)
  1. Response: ΔCDAI > 70 points from baseline.

  2. Remission: CDAI < 150 points.

Previous relapse frequency
 High31.129.158.30.03
 Low29.212.539.9
Steroids at induction
 Yes41.321.337.50.027
 No23.327.649.1
Combined immunosuppression at induction
 Yes5218300.001
 No23.327.449.3
Efficacy at week 12
 None10026.516.5< 0.001
 Response044.118.4
 Remission029.465.1
CRP < 10 mg/L at week 12 n = 189)
 Yes16.524.858.7< 0.001
 No50.727.922.1

In a logistic regression model, need for combined concomitant immunosuppression (coefficient: −1.44, OR: 0.24, 95%; CI: 0.07–0.81, P = 0.022) at induction, previous operations (coefficient: −0.95, OR: 0.39, 95%; CI: 0.15–0.99, P = 0.049), low CRP at week 12 (coefficient: 2.02, OR: 7.55, 95%; CI: 2.55–22.4, P < 0.001) and clinical efficacy at week 12 (coefficient: 3.63, P < 0.001) were identified as independent predictors for week 24 clinical remission after also having considered induction dose, frequency of relapses, perianal disease and smoking status in the analysis.

At week 52, frequency of previous relapses (P = 0.03), need for concomitant steroids (P = 0.03), combined immunosuppression (P = 0.001), low CRP at week 12 (< 10 mg/L, P < 0.001) and clinical efficacy (response or remission) at week 12 (P < 0.0001) were associated with clinical efficacy (Table 3). Clinical response at week 12 (P < 0.001), low CRP at week 12 (P < 0.001, OR: 5.02, 95% CI: 2.55–9.88), need for combined concomitant immunosuppression at induction (P = 0.018, OR: 0.44, 95% CI: 0.22–0.88), frequency of previous relapses (P = 0.025, OR: 0.47, 95% CI: 0.21–0.92) smoking status (P = 0.07, OR: 0.52, 95% CI: 0.25–1.07) and shorter disease duration (< 3 years; P = 0.07, OR: 1.88, 95% CI: 0.94–3.74) were associated with 52-week clinical remission. Clinical remission (47.1% vs. 42.3%) or response (73.1% vs. 66%) rates at 12 months were numerically greater in patients naïve to infliximab, compared with patients previously exposed; however, this difference was not statistically significant. Similarly, remission rates in patients with parallel fistulising disease were numerically lower at each time-point assessed (remission fistulising vs. luminal only week 12: 42.5% vs. 56.3%, P = 0.06, week 24: 45.2% vs. 54.7% and week 52: 41.1% vs. 48.7%), although at no time-point, the differences were statistically significant.

In a logistic regression model, clinical efficacy at week 12 (response or remission) (P < 0.001), need for combined concomitant immunosuppression (P = 0.021) at induction, short disease duration (P = 0.03) and smoking (P = 0.049) were identified as independent predictors for week 52 clinical remission. If low CRP levels are included in the same analysis, thereby excluding 12 patients from analysis, the effect of smoking became borderline-significant (Table 4). Low CRP at week 12 was identified as an independent predictor.

Table 4.   Logistic regression: predictive factors for clinical remission during adalimumab therapy at 52 weeks in Crohn’s disease
FactorCoefficientP valueOR95% CI
  1. The coefficient is equivalent to the natural log of the OR; P value: level of significance; OR: odds ratio; 95% CI: 95% confidence interval.

  2. * < 10mg/L, † < 3 years.

Concomitant combined IS at induction−0.910.0440.400.17–0.97
Clinical efficacy at week 121.71< 0.0015.532.83–10.8
Low CRP at week 12*1.53< 0.0014.612.06–9.56
Short disease duration†1.140.0133.121.27–7.68
Smoking−0930.0550.390.15–1.02
High relapse frequency−0.680.11

In a further logistic regression model, we included only patients with a baseline elevated CRP. In this subgroup analysis (n = 131), normalised CRP (P = 0.006, OR: 3.72, 95% CI: 1.45–9.5) and clinical efficacy at week 12 (P < 0.0001, OR: 6.32, 95%; CI: 2.57–15.5) were identified as independent predictors of week 52 clinical remission. Table 6 shows the predictive power of the individual markers in distinguishing between patients reaching and not reaching clinical remission by week 52. Normal or normalised CRP at week 12 and clinical remission at either week 12 or week 24 were associated with the best accuracy.

Gender, location, behaviour, perianal disease, presence of extraintestinal manifestations, CRP at the start of the biological therapy, previous anti-TNF therapy or induction dose were not associated with clinical response or remission at either week 24 or week 52.

Need for dose escalation and predictors of dose escalation

Dose escalation to weekly dosing was needed in 16.4% of patients during the 1-year adalimumab therapy. Parallel AZA therapy was inversely associated with escalation to weekly dosing (P = 0.005, OR: 0.34, 95% CI: 0.16–0.74) and in a Kaplan–Meier analysis to time to escalation to weekly dosing (pLogRank = 0.003, pBreslow = 0.002).

Similarly, clinical remission at week 12 (pLogRank = 0.009, pBreslow = 0.004) or low CRP at week 12 (pLogRank = 0.026, pBreslow = 0.038) was associated with time to escalation to weekly dosing in a Kaplan–Meier analysis (Figure 1). In a Cox-regression analysis, concomitant AZA therapy (P = 0.018, HR: 0.41, 95% CI: 0.20–0.86), clinical remission at week 12 (P = 0.021, HR: 0.39, 95% CI: 0.18–0.87) but not CRP (P = 0.16) at week 12 were independently associated with the probability of dose escalation.

Figure 1.

 Association between azathioprine therapy, early clinical remission and need for dose escalation during the first year of adalimumab therapy. (a) Azathioprine (AZA), pLogRank = 0.003, pBreslow = 0.002. (b) pLogRank = 0.009, pBreslow = 0.004.

Mucosal healing and predictors of mucosal healing

Endoscopic partial healing and complete healing were achieved in 43.1% and 23.6% of CD patients, respectively, with available endoscopy at week 52 or dropout due to clinical deterioration or surgery (n = 123). Patients with shorter disease duration (< 3 years: n = 22) had numerically higher complete (27.3% vs. 22.8%) or partial (54.5% vs. 40.6%) mucosal healing rates, although the difference was not statistically significant.

Frequency of previous relapses (P = 0.040, OR: 0.45, 95% CI: 0.20–0.97), luminal only disease as indication for treatment (P = 0.007, OR: 2.91, 95% CI: 1.33–6.35), low CRP at week 12 (P < 0.001, OR: 8.76, 95%; CI: 3.44–22.3) and being in clinical remission at week 12 (P = 0.002, OR: 3.22, 95%; CI: 1.53–6.78) or at week 24 (P < 0.001, OR: 5.78, 95%; CI: 2.64–12.6) and tendential previous IFX therapy (P = 0.06, OR: 0.50, 95% CI: 0.24–1.03) were associated with endoscopic improvement/healing at 12 months.

In a logistic regression model, luminal only disease as indication for treatment, low CRP at week 12, being in clinical remission at week 24, and frequency of previous relapses and smoking were associated with endoscopic improvement and healing at 12 months (Table 5). Table 6 shows the predictive power of the individual markers in distinguishing between patients reaching and not reaching endoscopic improvement by week 52. Normalised CRP at week 12 was associated with the best accuracy.

Table 5.   Logistic regression: predictive factors for endoscopic response during adalimumab therapy at 52 weeks in Crohn’s disease
FactorCoefficientP valueOR95% CI
  1. The coefficient is equivalent to the natural log of the OR; P value: level of significance; OR: odds ratio; 95% CI: 95% confidence interval.

  2. * < 10mg/L.

Luminal only disease1.360.0083.891.43–10.6
Clinical remission at week 241.450.0034.251.64–10.9
Low CRP at week 12*1.92< 0.0016.842.34–20.0
High relapse frequency−1.000.0570.370.13–1.03
Smoking−1.330.0560.260.07–1.04
Table 6.   Predictive power of clinical and laboratory parameters to identify patients in clinical remission (CDAI < 150) (A) and endoscopic response (B) at week 52
 Sensitivity (%)Specificity (%)PPV (%)NPV (%)
  1. PPV, positive predictive value; NPV, negative predictive value.

  2. * Cut-off for CRP 10 mg/L.

  3. † Elevated CRP at diagnosis decreased to normal at week 12 (n = 122).

  4. ‡ Elevated CRP at diagnosis decreased to normal at week 12 (n = 67).

A
 Normal CRP week 12*83526375
 Normalised CRP at week 12†77626773
 Clinical remission at week 1275687073
 Clinical remission at week 2486777984
B
 Normal CRP week 12*86586781
 Normalised CRP at week 12‡80797980
 Clinical remission at week 1264666665
 Clinical remission at week 2470717170

Adverse events

Adverse events were monitored during the follow-up period. Eleven patients required major abdominal surgery (5.5%) and an additional six patients underwent perianal procedures, mainly due to abscess formation. Infectious adverse events developed in 12 patients (6.0%), including two cases of severe pneumonia (requiring hospitalisation), one new case of TB, one of CMV colitis, one of campylobacter gastroenteritis, two cases of herpes zoster, three skin infections, one instance of oral candidiasis and three other infections (recurrent bronchitis and gastroenteritis). Two previously asymptomatic patients reported development of psoriasis and another two presented with new arthritis symptoms despite previously being in clinical remission confirmed by a specialist. One patient developed an SLE-like syndrome. Two nonresponder patients experienced deep venous thromboses. One patient presented with allergic reaction. Eight patients reported various injection site reactions (4.0%).

Discussion

Major finding of this prospective follow-up study is that it was possible to identify clinically relevant patterns for the prediction of clinical and endoscopic response during medium-term adalimumab therapy. Clinical efficacy and CRP at week 12, need for combined immunosuppression, luminal disease and smoking status are predictors for medium-term clinical efficacy and mucosal healing during adalimumab therapy, whereas parallel AZA therapy may decrease the probability for dose escalation.

Overall clinical response and remission rates at week 24 were 78% and 52%, respectively; at week 52, clinical response and remission rates were 69.4% and 44.4%, respectively. Clinical remission rate of steroids was 38.8% at week 52. This is in concordance with results from previous clinical trials, which reported rates between 40% and 47%5–7 with a significant proportion of patients remaining free of abdominal surgery.10 Of note, the previous anti-TNF exposure rate of our patients (48.3%) was in the range reported in the randomised clinical trials with adalimumab. The rate of concomitant steroid therapy was 41.3% in concordance with the previous randomised clinical trials; however, the rate of concomitant AZA therapy was higher in the present patient population (69.2%) and none of the patients stopped AZA during the adalimumab therapy. The rate of triple immunosuppression was 26.4%. Moreover, patient selection was strict, due to the limited availability of the biological therapy, specialists took appropriate measures to objectively identify the underlying inflammatory burden. As a consequence, the rate of elevated CRP (> 10 mg/L) was 66.8%, and 80.1% of the patients had available endoscopy showing active inflammation within 1 year of the adalimumab therapy (54.5% severe endoscopic activity with deep ulceration).

The rate of parallel perianal or penetrating disease phenotype was high, and perianal fistulising disease was the primary or co-primary indication for the treatment in 36.3% of patients. Remission rates in patients with parallel fistulising disease were numerically lower at each time-point assessed in the present study, although statistical significance was not reached. The remission rate in fistulising disease in the present study was in the range reported in previous clinical trials with adalimumab (41% in CHOICE).11 In concordance, a higher efficacy of infliximab induction was reported in patients with luminal disease in a short-term Italian multicentric trial.25

Mucosal healing is becoming an increasingly important factor for measuring treatment success, and it is hoped that achieving mucosal healing will lead to improved long-term outcomes, as demonstrated during the follow-up of the step-up/top-down trial.26 In this study, complete mucosal healing, defined as a simple endoscopic score of 0 after 2 years of therapy, was the only factor that predicted sustained, steroid-free remission 3 and 4 years after therapy was initiated. However, only a minority of the original patient group has undergone endoscopic re-evaluation. The only study that had an endoscopic primary endpoint was the EXTend the Safety and Efficacy of Adalimumab Through ENDoscopic Healing (EXTEND), which evaluated the efficacy of adalimumab therapy in achieving deep remission (CDAI < 150 and absence of ulcers).27 Moreover, the mucosal healing rate was associated with disease duration (< 2 years: 44%, 2–4 years: 40% vs. > 5 years: 21%). In the present study, endoscopic partial and complete (absence of ulcers) healing were achieved in 43.1% and 23.6% in CD patients with available endoscopy at week 52 or dropout due to clinical deterioration or surgery (n = 123). In addition, we did not find any significant association between presence of complete mucosal healing and disease duration, although patients with shorter disease duration (< 3 years: n = 22) had numerically higher complete or partial mucosal healing rates.

In the present study, we identified a combination of clinical and laboratory markers predictive of medium-term clinical efficacy. Similarly, in an article by Leuven group,8 the predictors of the long-term outcome of the IFX treatment were analysed in CD patients. Initial response to IFX and sustained clinical benefit of IFX up to the end of follow-up were associated with a significant decrease in the need for major abdominal surgery (MAS). In addition, a decrease in elevated baseline CRP levels back to normal values (< 3 mg/L) after initiation of IFX treatment was significantly associated with a better outcome of IFX treatment in the long term.8 Of note, the present study used a different cut-off for CRP (< 10 mg/L) based on randomised clinical trials with adalimumab6 and certolizumab10, 11 and for results from our earlier study.28 Furthermore, in a phase 2 study,11 a greater difference in response and remission rates was reported in a post hoc analyses between the certolizumab group and the placebo group in patients with a baseline CRP level exceeding 10 mg/L. In the present study, normal/normalised CRP at week 12, but not baseline CRP was associated with medium-term clinical benefit. Similarly, normal/normalised CRP at week 12, clinical remission at week 24, frequency of previous relapses and smoking were associated with endoscopic improvement and healing. The predictive power of the markers was also acceptable. In addition, to enable a better comparison with the Leuven data, we also calculated the predictive power of normalised CRP at week 12 using the upper limit of the normal as cut-off (5 mg/L, n = 155). Nonetheless, the accuracy to predict clinical (sensitivity: 54%, specificity: 62%, PPV: 59% and NPV: 59%) or endoscopic remission (sensitivity: 57%, specificity: 78%, PPV: 72% and NPV: 64%) was inferior compared with higher cut-off (10 mg/L) proposed by the present study.

Smoking is an established risk factor for relapses, surgery and need for immunosuppressants in Crohn’s disease.1 There is increasing evidence that smoking may affect the efficacy of anti-TNFs and may increase the risk of loss of response. The results on the association between smoking and infliximab efficacy published so far are conflictive.29–32 In the present study, smoking was identified as a predictor of adverse clinical and endoscopic outcome at 12 months. The statistical significance was borderline in multiple regression analysis. A recent study on 30 patients treated with adalimumab also suggested that smokers are significantly less likely to respond to treatment and that, this effect correlates with the number of cigarettes smoked per day.33

Previous infliximab nonresponse and previous treatment with an anti-TNF agent were also reported to be associated with decreased clinical efficacy and loss of response.12, 34 In the present study, clinical remission (47.1% vs. 42.3%) or response (73.1% vs. 66%) rates at 12 months was numerically greater in patients not exposed previously to infliximab; however, the difference was not statistically significant. Nonetheless, the reason for infliximab discontinuation included numerous clinical scenarios from primary nonresponse to induction only or episodic therapy. The relative importance of the different causes could not be studied due to the small number of patients in the different subgroups. However, the remission rate for previous infliximab primary nonresponders was only 33.3%.

Finally, dose intensification to weekly dosing was needed in 16.4% of patients, and parallel AZA therapy and clinical remission at week 12 were inversely associated with dose escalation to weekly dosing in the present study. Corresponding rates were higher in the CHARM trial. Of the 260 patients randomised to adalimumab every other week (starting maintenance dose), 27.3% changed to weekly dosing during year 1 and an additional 13.1% changed to weekly dosing during year 2.35 In a recent meta-analysis,33 the mean percentage of patients who needed an adalimumab dose escalation was 21.4% and the annual risk was 24.4% per patient-year, when both initial responders and patients with primary nonresponse were analysed. Similar to the findings of this study, in a previous study by the Leuven group,13 time to dose escalation during adalimumab therapy was longer in patients resistant to infliximab who were treated with concomitant immunomodulators. In contrast, the importance of concomitant immunosuppression was not proven for adalimumab in post hoc analysis of the randomised controlled trials.6 Similarly, cessation of concomitant immunosuppressants did not translate to decreased clinical efficacy during a 24-month follow-up when immunosuppressants were either discontinued or continued after 6 months.36 Nevertheless, new data on combination therapy with immunosuppressants and infliximab have shown increased efficacy as compared with biological-only therapy in both AZA naïve and AZA-exposed CD patients.17, 18

The induction regimen also appeared to impact long-term efficacy, at least in some studies. Loftus et al.37 found that patients receiving an induction regimen of 160/80 mg were half as likely to receive weekly dosing because of loss of response compared with 80/40 mg regimen. Regional variation in the United States population was also described. In contrast, we were unable to confirm the above association in the present study. Of note, however, patients in the present study were not randomised to the induction regimes. Instead, the induction dose was selected according to clinical decision by the treating gastroenterologist.

The authors are aware of possible limitations of this study. Drug-trough levels and antibody status were not assessed. In the study by the Leuven group,13 besides CRP, drug-trough level was identified as a predictor for sustained clinical benefit during a median 20-month adalimumab therapy in CD patients refractory to infliximab. Significantly lower adalimumab trough serum concentrations were measured throughout the follow-up period in patients who discontinued therapy as compared with patients who stayed on adalimumab. Antibodies against adalimumab were present in 9.2% of patients and there was a corresponding effect on trough serum concentration. However, this is not representative of scenarios encountered in everyday clinical practice, as the patient population selected for this study was a challenging group resistant to infliximab. Thus, although the value of the serum drug concentration and antibody determination is unquestionable when patients lose response,33 the routine determination and predictive value of either drug-trough levels or antibodies in addition to stringent clinical follow-up is questionable. At least lower infliximab drug concentration did not translate into reduced clinical efficacy during a 24-month follow-up when immunosuppressives were either discontinued or continued after 6 months.36

In contrast, one of the strengths of the present study is that we included a cohort of well-characterised CD patients representing the nationwide clinical practice in Hungary with standardised patient selection and follow-up, according to the regulations of the Hungarian Insurance Bureau (OEP). Moreover, all centres were monitored for quality of care and compliance with the regulations by the OEP in June 2011. In addition, a large number of patients were also available for endoscopic assessment preceding and following the anti-TNF therapy.

In conclusion, results from this multicentre study suggest that early clinical efficacy, normal/normalised CRP at week 12 (early deep clinical remission) and luminal disease are associated with medium-term clinical efficacy and mucosal healing during adalimumab therapy, whereas need for combined immunosuppression at induction and tendencially smoking are predictors for non-response. Parallel azathioprine therapy may decrease the probability for dose escalation. In contrast, gender, presence of extraintestinal manifestations, CRP at the start of the biological therapy, previous anti-TNF therapy or induction dose were not associated with clinical response or remission at either week 24 or week 52.

Acknowledgements

Declaration of personal interests: The authors thank Michael D Mandel for reviewing and editing the manuscript. Drs. Aron Vincze, Karoly Palatka, Beata Gasztonyi and Gabor Horvath have served as speakers and advisory board members for Schering-Plough/MSD. Drs. Tamas Molnar, Pal Miheller and Ferenc Nagy have served as speakers and advisory board members for Schering-Plough/MSD and Abbott Laboratories. Dr. Peter Laszlo Lakatos has served as a speaker, a consultant and an advisory board member for Schering-Plough/MSD and Abbott Laboratories, and has received unrestricted research funding from Schering-Plough/MSD. Declaration of funding interests: None.

Ancillary