Longterm safety and efficacy of certolizumab ... - Wiley Online Library

Alimentary Pharmacology and Therapeutics

Long-term safety and efficacy of certolizumab pegol in the treatment of Crohn’s disease: 7-year results from the PRECiSE 3 study W. J. Sandborn*, S. D. Lee†, C. Randall‡, A. Gutierrez§, D. A. Schwartz¶, S. Ambarkhane**, C. Kayhan††, B. Pierre-Louis‡‡, S. Schreiber§§ & G. R. Lichtenstein¶¶

*Division of Gastroenterology, University of California San Diego School of Medicine, La Jolla, CA, USA. † Department of Medicine, Clinical Inflammatory Bowel Disease Program, University of Washington School of Medicine, Seattle, WA, USA. ‡ Department of Medicine, University of Texas at San Antonio, Gastroenterology Research of America, San Antonio, TX, USA. § Division of Gastroenterology, Washington University in St. Louis School of Medicine, St. Louis, MO, USA. ¶ Inflammatory Bowel Disease Center, Department of Medicine, Vanderbilt University Medical Center, Nashville, TN, USA. **UCB Pharma, Monheim, Germany. †† TA Immunology, UCB Pharma, Raleigh, NC, USA. ‡‡ UCB Pharma, Raleigh, NC, USA. §§ Department of Internal Medicine and Institute for Clinical Molecular Biology, Christian-Albrechts University at Kiel, Kiel, Germany. ¶¶ Center for Inflammatory Bowel Disease, Division of Gastroenterology and Hepatology, Perelman School of Medicine of the University of Pennsylvania, Philadelphia, PA, USA.

Correspondence to: Dr W. J. Sandborn, Division of Gastroenterology, University of California San Diego, 9500 Gilman Drive, La Jolla, CA 92093-0956 USA. E-mail: [email protected] Publication data Submitted 16 May 2014 First decision 28 May 2014 Resubmitted 30 July 2014 Accepted 31 July 2014 EV Pub Online 22 August 2014 This article was accepted for publication after full peer-review.

ª 2014 John Wiley & Sons Ltd doi:10.1111/apt.12930

SUMMARY Background The efficacy and safety of certolizumab pegol (CZP) in moderate-to-severe Crohn’s disease were demonstrated in two 26-week double-blind studies (PRECiSE 1 & 2). Aim To report the safety and efficacy outcomes of long-term, CZP therapy from PRECiSE 3, in which patients received treatment up to 7 years treatment. Methods Patients completing PRECiSE 1 or 2 were eligible to enter PRECiSE 3 in which they received CZP 400 mg, open-label, every 4 weeks (without additional induction therapy) for up to 7 years, for up to 91 doses from study start. Safety (adverse events, including infections and malignancies) and efficacy (Harvey–Bradshaw Index, faecal calprotectin, C-reactive protein) were prospectively monitored. Remission was analysed using observed cases, last observation carried forward imputation and nonresponder imputation. Results A total of 595 patients entered the study; 117 (20%) completed 7 years. Discontinuation rates were 29.2%, 13.6%, 16.1%, 7.9%, 5.0%, 4.5% and 3.9% (years 1–7 respectively). During 1920 patient-years of exposure to CZP, no new safety signals were observed. Incidence rates (new cases/100 patientyears) for serious infections and malignant neoplasms were 4.37 and 1.06 respectively. No lymphoproliferative malignancies were reported. Clinical remission rates were ≥68% at each year (observed cases); rates by last observation carried forward and nonresponder imputation were 58% and 45% at year 1, 56% and 26% at year 3 and 55% and 13% at year 7 respectively. Conclusion Certolizumab pegol was well tolerated in the long-term treatment of Crohn’s disease, with sustained remission in some patients continuing in the study for up to 7 years. ClinicalTrials.gov identifier NCT00552058. Aliment Pharmacol Ther 2014; 40: 903–916

903

W. J. Sandborn et al. INTRODUCTION Crohn’s disease (CD) is an incurable chronic inflammatory disease of the gastrointestinal tract. Symptoms include diarrhoea, abdominal pain and weight loss.1 Treatments are aimed at suppressing inflammation and inducing remission. Therapies for CD include corticosteroids, immunosuppressive agents such as azathioprine and mercaptopurine, and metabolite inhibitors such as methotrexate. Most patients treated with these agents experience disease progression and require biologic therapies or surgical resection of the affected bowel during the course of their disease.2 Biologic therapies such as tumour necrosis factor (TNF) a antagonists and antiintegrin antibodies have made clinical remission common and an achievable target of treatment. Over the last decade, TNF antagonists including infliximab, adalimumab and certolizumab pegol have become widely used for the treatment of CD. All of these agents are approved for the treatment of moderate-to-severe CD in adults to induce clinical response and/or remission.3 To date, observational and retrospective data have been analysed for safety and efficacy for treatment periods up to 10 years4–7 and efficacy data for prospective clinical trials have been reported for periods of up to 4 years.8–10 However, the durability of remission and tolerance of patients to anti-TNF agents remains an important question in clinical practice. In particular, it is unknown whether the risk of adverse events (AE) such as serious and opportunistic infections;11, 12 malignancies, including lymphomas, melanoma and nonmelanoma skin cancer;13 demyelinating disorders; lupus-like syndrome; congestive heart failure; new onset/worsening of psoriasis and death,14, 15 which have been associated with anti-TNF antibodies, increase with length of exposure to anti-TNF agents. However, long-term controlled clinical trials are challenged by patient discontinuation. Common reasons for discontinuation include lack (or loss) of efficacy, AEs, patient decision (often for personal reasons such as transportation, inconvenience or relocation), physician decision (for reasons other than lack of efficacy or AEs), sponsor decision (the sponsor may decide that a patient should be discontinued following consultation with the investigator) and loss to follow-up (the patient did not attend a scheduled visit and could not be contacted).16 Certolizumab pegol (CZP) is a humanised, PEGylated, anti-TNFa monoclonal antibody with high affinity for TNF and is administered subcutaneously. CZP is Fc-free and therefore does not display Fc-mediated cellular cytotoxicity.17 Two phase III, multicentre,

904

placebo-controlled, double-blind, randomised trials (PRECiSE 118 and PRECiSE 219) demonstrated that CZP is effective for induction and maintenance of clinical response in patients with moderate to severely active CD. In these trials, the study drug was administered as a 6-week treatment induction [CZP 400 mg (PRECiSE 1 and 2) or placebo (PBO, PRECiSE 1) at weeks 0, 2 and 4] followed by an additional 20 weeks of maintenance therapy (CZP 400 mg or PBO every 4 weeks). To evaluate the long-term safety and efficacy of CZP for moderate-to-severe CD, a long-term, openlabel extension of PRECiSE 1 and 2 was conducted (PRECiSE 3), in which patients who completed the double-blind trials were treated with open-label CZP 400 mg every 4 weeks.20 The induction regimen was not administered during the extension study regardless of the initial treatment that patients received. A second open-label extension study was conducted for patients who experienced exacerbation of CD and were not able to complete a 26-treatment period of the two pivotal trials (PRECiSE 421); details of this study will be reported in a subsequent manuscript. In this manuscript, we report 7-year safety and efficacy outcomes from PRECiSE 3. This is the first report of safety and efficacy data up to 7 years from a prospective clinical study of an anti-TNF agent for treatment of moderately to severely active CD.

METHODS Patients Patients who completed either of the PRECiSE 1 or 2 studies (which began in December 2003 and February 2004 respectively18, 19) were eligible for enrolment in the long-term extension study, PRECiSE 3 (ClinicalTrials.gov identifier NCT00160524), which continued until August 2012. For eligible patients, enrolment into the PRECiSE 3 study was based on individual patient decision and did not involve the study sponsor. Enrolment criteria for PRECiSE 1 and 2 (initial studies) have been previously described.18, 19 Briefly, patients were ≥18 years of age and diagnosed with moderate-to-severe CD [Crohn’s Disease Activity Index (CDAI) of 220–450 points scored over the week prior to the first study dose, suggesting a failure of conventional treatments]. Approximately 20% of patients were previously treated with, and had responded to, infliximab.18, 19 For those who had been previously treated with infliximab, primary response was required for entry into PRECiSE 1 and 2.

Aliment Pharmacol Ther 2014; 40: 903-916 ª 2014 John Wiley & Sons Ltd

7-year data of certolizumab pegol in Crohn’s disease Patients with short-bowel syndrome, obstructive symptoms, or evidence of recent tuberculosis were excluded. Patients could not have received any TNF antagonist within 3 months prior to the initial studies or have severe hypersensitivity to any TNF antagonist.

Study design The designs for the two initial studies have been previously described.18, 19 Briefly, at the beginning of the trial patients in PRECiSE 1 were randomised to receive induction therapy (PBO or CZP 400 mg at weeks 0, 2 and 4), followed by maintenance therapy with the same study drug (PBO or CZP 400 mg every 4 weeks up to week 26). All patients in PRECiSE 2 were administered induction therapy with CZP 400 mg (at weeks 0, 2 and 4). At week 6, patients who responded to therapy were randomised to receive PBO or CZP 400 mg every 4 weeks up to week 26. PRECiSE 3 was a phase III, open-label extension study to assess the effects of CZP 400 mg given every 4 weeks up to week 362 (7 years) to patients with CD who completed the initial studies. Patients who received PBO during the initial studies were also included and treated with CZP. In PRECiSE 3, patients could have received up to 91 doses of CZP and were assessed on up to 94 visits. Dose escalation and dose-splitting were not permitted at any point during the study. The extension study was conducted in 29 countries in the regions of North America (Canada and the United States), Europe (Austria, Belarus, Belgium, Bulgaria, the Czech Republic, Denmark, Estonia, Georgia, Germany, Hungary, Italy, Latvia, Lithuania, Norway, Poland, Russia, Serbia, Slovenia, Spain, Sweden and Ukraine) and the rest of the world (Australia, Hong Kong, Israel, New Zealand, Singapore and South Africa). All nonstudy drugs being taken at entry into PRECiSE 3 had to remain at or below a stable dose throughout the duration of the study. These included 5-aminosalicylates, prednisolone at a dose of ≤30 mg/ day, azathioprine, mercaptopurine, methotrexate or antibiotics. However, tapering of steroids or immunosuppressants was allowed and was based upon investigator discretion. Patients were considered to have treatment failure if they required rescue therapy. Rescue therapy was defined as the patient needing treatment for exacerbation of their CD with any of the following: TNF antagonists (other than CZP), increased doses of corticosteroids (increase of >15 mg/day prednisolone, or equivalent), increased doses of immunosuppressive agents, surgery and in-patient Aliment Pharmacol Ther 2014; 40: 903-916 ª 2014 John Wiley & Sons Ltd

hospitalisation. Any data collected after the first administration of rescue medication were not included in the efficacy analyses and those patients were classified as nonremitters from and including the time of receiving the rescue therapy. For patients already receiving corticosteroids increases of >15 mg/day prednisolone or equivalent or any increases in immunosuppressants, in response to exacerbation of their CD were regarded as rescue therapy (and patients were thus classified as treatment failure and withdrawn from the study). The PRECiSE studies were conducted in compliance with the principles of the Declaration of Helsinki and the International Conference on Harmonisation Good Clinical Practice guidelines. In addition, all local regulatory requirements were followed. The protocol was approved by the Institutional Review Board or Ethics Committee at each centre. All patients signed an informed consent form.

Assessments Patients were assessed for safety and efficacy at each visit during the study for up to 94 visits. The primary objective of the PRECiSE 3 study was assessment of the safety of long-term exposure to CZP. Safety was evaluated by observation and direct questioning at each visit, in addition to lab parameters, chest X-ray and physical exam at select visits. In order to ensure complete safety data collection, a safety follow-up visit was conducted 12 weeks after the last dose. AEs and serious AEs (SAE) were assessed, classified and descriptively summarised by primary system organ class and preferred term according to the Medical Dictionary for Regulatory Activities (MedDRA version 6.1). A secondary objective of this study was to assess efficacy during long-term treatment with CZP for CD. In PRECiSE 1 and 2, data for CDAI and the Harvey– Bradshaw Index (HBI) were collected, and the correlation between two indexes was established.22 In PRECiSE 3, efficacy was assessed using the HBI.23, 24 The HBI is scored by a sum of five parameters: general well-being (0–4; 0 = very well, 4 = terrible), abdominal pain (0–3; 0 = none, 3 = severe), number of liquid stools per day, abdominal mass (0–3; 0 = none, 3 = definite and tender) and complications (1 per item: arthralgia, uveitis, erythema nodosum, aphthous ulcers, pyoderma gangrenosum, anal fissure, new fistula and abscess). Clinical remission was defined as an HBI score of ≤4 points. CRP was measured every 4 weeks up to week 14, at weeks 26 and 54, then approximately every 52 weeks 905

W. J. Sandborn et al. until the end of the study. Plasma concentrations of CZP and anti-CZP antibodies were assessed and will be reported in a separate publication.

Statistical methods The study population described in the present analysis includes all patients who received maintenance treatment with CZP or PBO in PRECiSE 1 and 2 and who continued with CZP maintenance treatment in PRECiSE 3. All data in this report were analysed and are presented in a descriptive manner; no statistical significance testing was performed. As safety was the primary objective, this study was not powered for statistical hypothesis testing of efficacy end points. Patients were divided into three categories based on their randomised treatment regimens during the initial

studies (Figure 1): those randomised to PBO in PRECiSE 1 for 26 weeks were categorised as ‘First exposure group’ (n = 166), those treated with CZP induction (through week 6) followed by randomisation to PBO for 20 weeks in PRECiSE 2 were categorised as ‘Re-exposure group’ (n = 100) and those treated continuously with CZP (PRECiSE 1 and PRECiSE 2) were categorised as ‘Continuous exposure group’ (n = 329). The times to patient discontinuation due to any reason, adverse events, and lack of improvement or disease deterioration were examined by Kaplan–Meier estimates for each of the three exposure groups. Rates of clinical remission were calculated using observed case (postbaseline scores where patients who withdrew or received rescue therapy or have missing efficacy data are not included in the denominator), last

PRECiSE 2 Eligible for PRECiSE 3 CZP-treated (n = 151) PBO-treated (n = 109) Enrolled in PRECiSE 3 CZP-treated (n = 141) PBO-treated (n = 100) Did not enroll (n = 19)

Enrolment

PRECiSE 1 Eligible for PRECiSE 3 CZP-treated (n = 202) PBO-treated (n = 176) Enrolled in PRECiSE 3 CZP-treated (n = 188) PBO-treated (n = 166) Did not enroll (n = 24)

Allocation

Re-exposure PRECiSE 2 PBO group (n = 100)

Continuous Exposure PRECiSE 1 and 2 CZP group (n = 329)

Overall (N = 595)

Discontinued treatment (n = 134) Adverse event (n = 61) Protocol noncompliance (n = 3) Patient decision (n = 45) Clinical decision (n = 19) Lost to follow-up (n = 1) Lack of improvement/ disease deterioration (n = 26) Other (n = 6)

Discontinued treatment (n = 81) Adverse event (n = 35) Protocol noncompliance (n = 4) Patient decision (n = 25) Clinical decision (n = 12) Lost to follow-up (n = 8)



Completed

First Exposure PRECiSE 1 PBO group (n = 166)

Discontinued

PRECiSE 3 (N = 595)

Completed (n = 32)

Lack of improvement/ disease deterioration (n = 17) Other (n = 10) Completed (n = 19)

Completed (n = 66)

Completed (n = 117)

Figure 1 | Patient disposition and classification by exposure to CZP during the 7-year PRECiSE 3 study. Note: more than one reason for discontinuation may have been recorded for a single patient. CZP, certolizumab pegol; PBO, placebo. 906

Aliment Pharmacol Ther 2014; 40: 903-916 ª 2014 John Wiley & Sons Ltd

7-year data of certolizumab pegol in Crohn’s disease observation carried forward (LOCF, remission rates are derived from the last observed postbaseline HBI scores carried forward) and nonremitter imputation (NRI, patients who received rescue therapy, withdrew or were lost to follow-up were classified as nonremitters from that time point onward) analyses. The patients were classified as nonremitters if they withdrew prior to completion of the study or received rescue medication. The nonremitters classification was applied from, and including, the time of withdrawal or introduction of rescue treatment. In addition, patients who did not have all the required data to derive a remission status were classified as nonremitters for that particular time point. The median CRP concentrations were calculated for the intent-to-treat population.

RESULTS Patient disposition A total of 638 subjects who completed the initial studies were eligible for inclusion in PRECiSE 3, which enrolled 595 patients (354 from PRECiSE 1 and 241 from PRECiSE 2). Twenty per cent (n = 117) of patients completed the 7-year study, while 478 gradually discontinued (45% due to AEs, 35% due to patient decision, 21% due to no improvement or disease deterioration, 13% due to clinical decision, 4% due to protocol noncompliance, 4% were lost to follow-up and 8% discontinued for other reasons (patients may have discontinued for more than one reason; Figure 1). Approximately 71% of patients were exposed to CZP for >1 year, and the mean number of CZP doses was 41. Demographics The baseline characteristics (prior to enrolment in PRECiSE 1 or 2 studies) of the patients are shown in Table 1. The mean (standard deviation) patient age was 38.1 (11.9) years, 52% of patients were female and 95% were white. The median duration of disease was 5 years. Ileocolonic location of disease was most common (46% of patients), as was inflammatory disease behaviour (68%). Prior resections were performed in 30% of patients. Overall, 20% of patients previously received infliximab therapy for CD but had developed intolerance or lost response.18, 19 At the start of the extension study, concomitant corticosteroids were used by 29% of patients and immunosuppressive agents were used by 35%. Demographic characteristics were generally similar across the first exposure, re-exposure and continuous exposure groups. Aliment Pharmacol Ther 2014; 40: 903-916 ª 2014 John Wiley & Sons Ltd

Safety Discontinuation rates by year of study in the overall population (N = 595) were 29.2%, 13.6%, 16.1%, 7.9%, 5.0%, 4.5% and 3.9% during years 1–7 respectively; the rates of discontinuation within each year were 29.2% (174/595, year 1), 19.2% (81/421, year 2), 28.2% (96/340, year 3), 19.3% (47/244, year 4), 15.2% (30/197, year 5), 16.2% (27/167, year 6) and 16.4% (23/140, year 7). Overall, AEs were the most common reason for patient discontinuation in PRECiSE 3, accounting for 44.6% (213/ 478) of the total withdrawals (patients could have discontinued for more than one reason, Figure 1). For the overall population, of those who discontinued during year 1, approximately half (84/174, 48.3%) did so due to AEs. In addition to the AEs, the most common reasons for discontinuation were patient decision and lack of improvement/disease deterioration; discontinuation for any reason, AEs and lack of improvement/disease deterioration were common during the first year of the PRECiSE 3 extension study, but tapered during subsequent years and rates were similar across the three exposure groups (Figure 2A–C). Discontinuation due to patient decision was also very common, but rates were not significantly different between the CZP exposure groups. Over 7 years of CZP treatment, and with 1920 patient-years at risk, 88.2% of patients in this study experienced an AE, and 40.3% experienced an SAE (Table 2). Rates of AEs, SAEs and infections were similar among the three exposure category groups and were relatively stable over time (Figure 3A–D). There were 204 AEs (34.3%) that resulted in patient’s withdrawal from the study. In the overall group, CD exacerbation was the most common AE (30.3%), the most common SAE (11.3%) and the most common AE leading to withdrawal (17.3%). The most common infection was nasopharyngitis (15.3%). Patients taking concomitant corticosteroids were more likely to experience a serious infection (58/ 366, 15.8%; overall safety population) than those not taking corticosteroids (21/229, 9.2%). If a patient reported more than one AE in a specific system organ class, that patient was only counted once for that class. Opportunistic infections were reported in 114 patients. The incidence rate for new serious infections was 4.37 cases/100 patient-years. New cases (diagnosed after screening of the initial studies) of pulmonary tuberculosis were reported in 3 patients (0.2 cases/100 patientyears; Table 3). Specifically, the patients included 2 females [age 31 years at screening, time from first injection of CZP (time) 1 year + 13 days, and age 47 years at screening, time 2 years + 3 months, both from South 907

W. J. Sandborn et al. Table 1 | Demographic characteristics were similar among the three exposure groups and the overall population at study baseline of PRECiSE 1 and PRECiSE 2, safety population

Age (years), mean (s.d.) Female, n (%) Race, n (%) White Body mass index, mean (s.d.) Disease duration from diagnosis (years), mean (s.d.) Location of CD, n (%) L1 – terminal ileum L2 – colon L3 – ileocolon L4 – upper GI* Behaviour of CD, n (%) B1 – inflammatory disease B2 – stricturing disease B3 – penetrating disease Prior infliximab use, n (%) Number of prior infliximab infusions, mean (s.d.) Corticosteroid use, n (%) Immunosuppressant use, n (%) Use of both corticosteroid and immunosuppressant, n (%) Prior resections, n (%) Number of resections, n (%)† 0 1 2 3 >3 Site of resections, n (%) Ileocolon Upper GI‡ Terminal Ileum Colon

First exposure group (n = 166)

Re-exposure group (n = 100)

Continuous exposure group (n = 329)

Overall safety population (N = 595)

39.0 (11.7) 93 (56.0)

37.2 (11.5) 45 (45.0)

37.9 (12.1) 169 (51.4)

38.1 (11.9) 307 (51.6)

156 (94.0) 24.2 (5.6) 7.5 (7.5)

90 (90.0) 24.7 (5.1) 6.9 (6.8)

319 (97.0) 24.1 (5.5) 7.2 (7.3)

565 (95.0) 24.2 (5.5) 7.3 (7.3)

41 36 82 7

(24.7) (21.7) (49.4) (4.2)

26 23 45 6

(26.0) (23.0) (45.0) (6.0)

76 96 147 10

(23.1) (29.2) (44.7) (3.0)

143 155 274 23

(24.0) (26.1) (46.1) (3.9)

108 26 32 29 6.4

(65.1) (15.7) (19.3) (17.5) (5.0)

71 8 21 17 5.2

(71.0) (8.0) (21.0) (17.0) (5.3)

224 43 62 73 5.0

(68.1) (13.1) (18.8) (22.2) (4.3)

403 77 115 119 5.4

(67.7) (12.9) (19.3) (20.0) (4.6)

58 (34.9) 52 (31.3) 21 (12.7)

22 (22.0) 34 (34.0) 9 (9.0)

95 (28.9) 124 (37.7) 36 (10.9)

175 (29.4) 210 (35.3) 66 (11.1)

60 (36.1)

31 (31.0)

90 (27.4)

181 (30.4)

106 44 12 3 1

(63.9) (26.5) (7.2) (1.8) (0.6)

69 19 7 3 2

(69.0) (19.0) (7.0) (3.0) (2.0)

239 62 18 3 7

(72.6) (18.8) (5.5) (0.9) (2.1)

414 125 37 9 10

(69.6) (21.0) (6.2) (1.5) (1.7)

38 1 25 5

(22.9) (0.6) (15.1) (3.0)

20 2 11 6

(20.0) (2.0) (11.0) (6.0)

52 2 36 12

(15.8) (0.6) (10.9) (3.6)

110 5 72 23

(18.5) (0.8) (12.1) (3.9)

CD, Crohn’s disease; GI, gastrointestinal tract; s.d., standard deviation. * If location is recorded as L4 – Upper GI, then terminal ileum, colon or ileocolon also must be recorded. † Patients may have had CD in more than one location. ‡ Upper GI includes all areas proximal to the terminal ileum and excludes the terminal ileum, ileocolon and colon.

Africa] and 1 male (age 56 years at screening, time 1 year + 3 months, from South Africa). Disseminated tuberculosis was reported in 3 patients (0.2 cases/100 patient-years), including 2 females (age 35 years at screening, time 1 year + 4 months, from Denmark and age 32 years at screening, time 1 year + 4 months, from South Africa) and one patient (female, age 47 years at screening, from South Africa) who died from disseminated tuberculosis after receiving 33 injections of CZP 908

(time 2 years + 6 months); this event was recorded as ‘definitely related to the study medication’ by the investigator. The five records of tuberculosis from South Africa originated from four different clinics. The incidence rate for malignant neoplasms was 0.84 cases/100 patient-years. Four patients experienced basal cell carcinoma (0.7%), one case of squamous cell carcinoma of the skin (0.2%) and one case of malignant melanoma (0.2%; Table 4). Of these 5 patients with Aliment Pharmacol Ther 2014; 40: 903-916 ª 2014 John Wiley & Sons Ltd

7-year data of certolizumab pegol in Crohn’s disease (a) 100

First exposure Re-exposure Continuous exposure Censored subjects - First exposure Censored subjects - Re-exposure Censored subjects - Continuous exposure

Patients not withdrawn (%)

90 80 70 60 50 40 30 20 10 0 0

50

100

150

200

250

300

350

400

Time to withdrawal (weeks) due to any reason


(b) 100 90 80 70 60 50 40 0

100

150

(c) 100

200

250

300

350

400


90

80

70 0

50

100

150

200

250

300

350

400

Time to withdrawal (weeks) due to lack of improvement or disease deterioration

nonmelanoma skin cancers, 1 was treated with concomitant azathioprine. No lymphoproliferative malignancies were reported in this study. New or worsening psoriasis was experienced by 11 patients (1.8%) in the overall population (incidence rate/ 100 patient-years: 0.58; 95% CI: 0.3, 1.0); the rates of Aliment Pharmacol Ther 2014; 40: 903-916 ª 2014 John Wiley & Sons Ltd

50

Time to withdrawal (weeks) due to adverse event


Figure 2 | Rates of discontinuation in the three exposure categories; Kaplan– Meier analyses of the most common reasons for discontinuation: any reason, adverse events, and lack of improvement or disease deterioration. (a) Discontinuation due to any reason. (b) Discontinuation due to adverse events. (c) Discontinuation due to lack of improvement or disease deterioration. (d) Discontinuation due to patient decision. Patients could have discontinued for more than one reason.


psoriasis across the three exposure groups were first exposure: 2.4%, re-exposure: 2% and continuous exposure: 1.5%. One patient in the continuous exposure group experienced psoriasis as an SAE. There were two cases of cardiac failure (0.3%), but neither was classified as congestive heart failure. There were no reports of 909

W. J. Sandborn et al. Table 2 | Adverse events were similar in the three exposure groups and the overall safety population over 7 years of treatment with CZP (N = 595) First exposure (n = 166) Any AE 150 (90.4) Any SAE 74 (44.6) Relationship to study drug* Unrelated 121 (72.9) Unlikely 113 (68.1) Possible 67 (40.4) Probable 15 (9.0) Definite 3 (1.8) Related to study drug† 75 (45.2) AE leading to death 0 AE leading to withdrawal 59 (35.5) Most frequent SAEs in the overall group by preferred term‡ Crohn’s disease 19 (11.4) Small intestinal obstruction 1 (0.6) Intestinal obstruction (unspecified) 5 (3.0) Serious infections in patients taking 15/112 (13.4) concomitant corticosteroids, n/N (%)‡,§ Serious infections in patients not taking 7/54 (13.0) concomitant corticosteroids, n/N (%)‡,§ Serious infections in patients with mild-moderate 9/62 (14.5) CD at baseline (HBI ≤8), n/N (%)‡ Serious infections in patients with moderate-severe 12/103 (11.7) CD at baseline (HBI >8), n/N (%)‡

Re-exposure (n = 100)

Continuous exposure (n = 329)

Overall (N = 595)

85 (85.0) 40 (40.0)

290 (88.1) 126 (38.3)

525 (88.2) 240 (40.3)

76 53 39 5 4 43 1 34

(76.0) (53.0) (39.0) (5.0) (4.0) (43.0) (1.0) (34.0)

230 187 125 20 11 134 3 111

(69.9) (56.8) (38.0) (6.1) (3.3) (40.7) (0.9) (33.7)

427 353 231 40 18 252 4 204

(71.8) (59.3) (38.8) (6.7) (3.0) (42.4) (0.7) (34.3)

9 3 1 11/58

(9.0) (3.0) (1.0) (19.0)

39 9 4 32/196

(11.9) (2.7) (1.2) (16.3)

67 13 10 58/366

(11.3) (2.2) (1.7) (15.8)

3/42 (7.1)

11/133 (8.3)

21/229 (9.2)

2/42 (4.8)

15/122 (12.3)

26/226 (11.5)

12/57 (21.1)

28/207 (13.5)

52/367 (14.2)

AE, adverse event; CD, Crohn’s disease; HBI, Harvey–Bradshaw Index; SAE, serious adverse event. Data are represented as n (%) unless otherwise indicated. * AEs with changing relationship to study drug over time are included only for the maximum reported relationship. † AEs with a possible, probable or definite relationship to study medication; any missing relationships are considered ‘related’. ‡ Where a patient experiences >1 AE in a category, the patient is only counted once in that category. § Concomitant corticosteroids, ongoing at or received after first injection in PRECiSE 3.

demyelinating disorders or lupus-like syndrome during this 7-year study. Overall, 4 patients (0.7%) died during the PRECiSE 3 study (Table 2). One patient experienced tracheal obstruction (aspiration of a piece of meat) after 14 injections of CZP. One patient experienced intestinal obstruction after receiving 11 injections of CZP. One patient experienced bacteriuria with uterine disorder and polycystic ovaries after receiving 50 injections of CZP. These first three events were considered unlikely to be related to the study medication by the investigators. A fourth patient died after experiencing disseminated tuberculosis as described above.

Efficacy During the PRECiSE 3 study, the remission rates (HBI ≤4) for the overall population as analysed by LOCF and 910

NRI methods were 58% and 45% at year 1, 56% and 26% at year 3 and 55% and 13% at year 7 respectively. By observed cases analyses, remission rates were 55% (325/591) at week 0 of the extension study and 75.7% (78/103) at year 7 for the patients who were still in the study at that time. A higher percentage of patients who began the extension study in clinical remission were in remission at all time points and according to all three analysis methods compared with those patients who were not in remission at Week 0. Remission rates for the first exposure group were generally lower than the re-exposure group or the continuous exposure group, and remained lower throughout the study (all three analysis methods; Figure 4). The median serum CRP concentrations in the overall population were 7 mg/L at baseline of the initial studies, 5 mg/L at the beginning of the extension study and Aliment Pharmacol Ther 2014; 40: 903-916 ª 2014 John Wiley & Sons Ltd

7-year data of certolizumab pegol in Crohn’s disease (a)

Total patients Adverse events

(c)

Serious adverse events Infections


Serious adverse events Infections 350

100

100 150

300

80

80

75

40

50 20

Rate (%)

Rate (%)

100

Population, n

60

250 200

40

150 100

20

25

0

60

0– 1. 1.5 5– 3 3– 6– 6 12 12 – 18 18 24 –24 – 30 30 36–36 – 42 42 –4 48 8 54–54 60–60 – 66 66 – 72 72 – 78 78 –8 >8 4 4

0– 1. 1.5 5– 3 3– 6– 6 12 12 – 18 18 24–24 – 30 30 36–36 – 42 42 – 48 48 54–54 60 –60 – 66 66 72 –72 – 78 78 –8 >8 4 4


100

80



100

600

80

500

25

20 0 0– 1. 1.5 5– 3 3– 6– 6 12 12 – 18 18 24–24 – 30 30 36–36 – 42 42 –4 48 8 54–54 60–60 – 66 66 72–72 – 78 78 –8 >84 4

0

Months

400 60 300 40 200 20

100 0

0 0– 1. 1.5 5– 3 3– 6– 6 12 12 – 18 18 24–24 – 30 30 – 36 36 – 42 42 – 48 48 54–54 60–60 – 66 66 72–72 – 78 78 –8 >84 4

50 40

Rate (%)

60

Population, n

75 Population, n

Rate (%)

(d)


100

0

Months

Months (b)

50

0

0

Population, n

125

Months

Figure 3 | Rates of adverse events, serious adverse events, infections by participating population were stable or declined over time in each of the exposure groups, safety population. (a) First exposure group (n = 166). (b) Reexposure group (n = 100). (c) Continuous exposure group (n = 329). (d) Overall population (N = 595).

5 mg/L, 4 mg/L, 3 mg/L, 3 mg/L, 3 mg/L, 3 mg/L and 2 mg/L at weeks 54, 106, 158, 210, 258, 310 and 364 respectively. Across the exposure groups, median CRP concentrations at baseline of the initial studies were similar (5 mg/L for First exposure, 8 mg/L for Re-exposure and 8 mg/L for Continuous exposure) and generally decreased over time during the extension study (Figure 5).

DISCUSSION The PRECiSE 3 study provided a unique opportunity to examine safety and efficacy of CZP treatment for CD over 7 years in a prospective phase III clinical trial. Previous prospective clinical trials for TNF antagonists in CD have demonstrated efficacy for up to 4 years,8–10 and Aliment Pharmacol Ther 2014; 40: 903-916 ª 2014 John Wiley & Sons Ltd

observational and retrospective safety and efficacy data have been reported for treatment periods up to 10 years.4–7 However, data regarding long-term patient outcomes of tolerance and efficacy in CD patients treated with anti-TNF therapy in a controlled setting remain limited. PRECiSE 3 is the longest prospective clinical trial to date that examines safety and efficacy of an anti-TNF for treatment of CD under the conditions of frequent standardised assessments and monitoring of patients. Long-term clinical trials may be difficult to complete for patients who experience a lack of disease improvement, AEs or who may have personal challenges with adhering to monthly visits to a clinic over multiple years. This study was able to retain participation of 20% 911

W. J. Sandborn et al. Table 3 | Frequency and incidence of clinically important infections over 7 years of exposure to CZP in the overall safety population (N = 595)

Preferred term

Overall population (N = 595) n (%)

Most frequent infections Nasopharyngitis 91 (15.3) Urinary tract infection 85 (14.3) Influenza 72 (12.1) Most frequent opportunistic infections Herpes zoster 16 (2.7) Fungal infection (unspecified) 13 (2.2) Tuberculosis Disseminated tuberculosis 3 (0.5) Pulmonary tuberculosis 3 (0.5) Most frequent serious infections Abdominal abscess 8 (1.3) Gastroenteritis 6 (1.0) Perianal abscess 9 (1.5)

Incidence of new cases per 100 patient-years (95% CI) 5.5 (4.4–6.7) 5.0 (4.0–6.2) 4.1 (3.2–5.2) 0.9 (0.5–1.4) 0.7 (0.4–1.2) 0.2 (0.0–0.5) 0.2 (0.0–0.5) 0.4 (0.2–0.8) 0.3 (0.1–0.7) 0.5 (0.2–0.9)

CI, confidence interval.

(n = 117) of patients through 7 years of treatment. An introspective evaluation of the patients who completed the study will be presented in a separate manuscript. The overall rates of discontinuation during this 7-year study were comparable to the 4-year CHARM extension study of adalimumab treatment for CD (ADHERE) (32% [n = 60/188] discontinued due to AEs, 20% [n = 37/ 188] discontinued due to loss of efficacy, 16% [n = 30/ 188] withdrew consent, 8% [n = 15/188] were lost to follow-up and 24% withdrew for other reasons.10 In a recent 5-year observational study of infliximab treatment for CD, reasons for discontinuations included loss of efficacy (n = 11/92, 12%), side effects (n = 12/92, 13%) and other reasons (n = 16/92, 17%).4 The rates of treatment continuation and compliance over multiple years may vary between clinical trials and observational or registry studies because of the strict visit schedule and enrolment/withdrawal requirements in the clinical trials versus the less restricted design applied to the observational studies. The highest percentage of patient discontinuation occurred during the first year of the PRECiSE 3 study (n = 174/595, 29%) and the primary reason for discontinuation was AEs. However, while the rates of AEs and SAEs remained relatively steady over the 7-year study, the rate of discontinuation due to AEs (number 912

Table 4 | Summary frequency and incidence of malignancies during 7 years of exposure to CZP in the overall safety population (N = 595)

Overall population, n (%)

System organ class and preferred term Any malignancies Basal cell carcinoma Mass Nodule Small intestine carcinoma Arachnoid cyst Dysplastic nevus syndrome Laryngeal cancer Lentigo maligna Lip neoplasm malignant Lung nodule Medullary thyroid cancer Metastatic malignant melanoma Nonsmall cell lung cancer Rectal cancer Squamous cell carcinoma of the skin Thyroid nodule

20 4 2 2 2 1 1 1 1 1 1 1 1

(3.4) (0.7) (0.3) (0.3) (0.3) (0.2) (0.2) (0.2) (0.2) (0.2) (0.2) (0.2) (0.2)

Incidence of new cases per 100 patient-years (95% CI) 1.06 0.21 0.10 0.10 0.10 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05

(0.6–1.6) (0.1–0.5) (0.0–0.4) (0.0–0.4) (0.0–0.4) (0.0–0.3) (0.0–0.3) (0.0–0.3) (0.0–0.3) (0.0–0.3) (0.0–0.3) (0.0–0.3) (0.0–0.3)

1 (0.2) 1 (0.2) 1 (0.2)

0.05 (0.0–0.3) 0.05 (0.0–0.3) 0.05 (0.0–0.3)

1 (0.2)

0.05 (0.0–0.3)

CI, confidence interval.

discontinued each year/population during the same year) declined rapidly after the first year. Discontinuations due to patient decision or lack of improvement/disease deterioration were reported nearly as often as AEs at each year of the study. Rates of discontinuation for any reason declined quickly and continuously after the first year of the study and were very low by year 7. No new safety signals were reported during this 7-year study that collected CZP data with 1920 patient-years at risk. No cases of demyelinating disorders, congestive heart failure or lupus-like syndrome were reported herein. The event rates of serious (including opportunistic) infections and malignancies were low over 7 years of treatment and were in line with the short-term studies18, 19 and with other long-term studies (adalimumab, serious infections, 6.1 events/100 patient-years and malignancies, 1.6 events/100 patient-years10). The incidence of tuberculosis in PRECiSE 3 was low overall and similar to data from adalimumab (0.2 events/ 100 patient-years10). Five of the six cases were reported in South Africa. It is worth noting that South Africa has one of the highest rates of tuberculosis in the world; the incidence rate of new tuberculosis was 1003 per 100 000 Aliment Pharmacol Ther 2014; 40: 903-916 ª 2014 John Wiley & Sons Ltd

0 12 3 4 56 7 012 345 6 7 01 2345 6 7 Years Years Years Re-exposure Continuous exposure First exposure

BL 0 1 2 3 4 5 6 7 BL0 1 2 3 4 5 6 7 BL0 1 2 3 4 5 6 7 Years Years Years Re-exposure Continuous exposure First exposure

75 50

44/329

89/329

73/329

56/329 51/329

157/329

117/329

18/100

16/100

193/329

29/100

25/100

18/100

69/100

41/100

52/100

19/165 18/165

34/165

28/165

48/165

27/165

25 63/165

57/165

0

189/329

191/329

185/329 188/329

189/329

193/329

193/329

66/100

203/329

66/100

68/100 66/100

69/100

69/100

69/100 68/100

70/165

71/165

74/165 73/165 75/165

75/165

75/165

25

50

75

Remission rate, NRI (%)

100

(c)

63/165

0

13/329

193/327 159/218 117/167 89/124 73/98 56/81 51/68 44/59

3/99 69/100 53/70 41/51 30/38 25/28 18/20 18/20 16/17

25

50

75

Remission rate, LOCF (%)

100

(b)

6/164 63/164 57/106 49/81 34/57 27/44 28/40 19/31 18/27

Remission rate, observed cases (%)

0

(a)

100

7-year data of certolizumab pegol in Crohn’s disease

012 3 4 56 7 012 345 6 7 01 2345 6 7 Years Years Years Re-exposure Continuous exposure First exposure

Median CRP concentration (mg/L)

Figure 4 | Rates of remission (HBI ≤4) by year in three different analyses, ITT populations. (a) Patients in remission (observed cases, postbaseline scores where patients who withdrew or received rescue therapy or have missing efficacy data are not included in the denominator). (b) Patients in remission from the last observation carried forward (LOCF, remission rates are derived from the last observed postbaseline HBI scores carried forward). (c) Patients in remission from nonremitters imputation (NRI, patients who received rescue therapy, withdrew or were lost to followup were classified as nonremitters from that time point onward). The error bars represent 95% CI for percentage. Years: BL = baseline of initial study, 0 = Week 0 (extension study), 1 = Week 54, 2 = Week 106, 3 = Week 158, 4 = Week 210, 5 = Week 258, 6 = Week 318, 7 = Week 364. HBI, Harvey–Bradshaw Index.

10 First exposure, n = 165 Re-exposure, n = 100 Continuous exposure, n = 329

8 6 4 2 0 0

50

100

150

200

250

300

350

400

Week

Figure 5 | Median CRP concentrations declined or remained stable over 7 years of treatment. CRP concentrations, mg/L serum. CRP was measured every 4 weeks up to week 14, at weeks 26 and 54, then approximately every 52 weeks until the end of the study. Week 0 is the beginning of the PRECiSE 3 extension study. CRP, C-reactive protein.

in 2012.25 The only other country where PRECiSE 3 was conducted with a high incidence rate is the Russian Federation, with 91 per 100 000. In contrast the incidence rate of tuberculosis in the United States was 3.6 per 100 000. These data support the need for screening and surveillance of tuberculosis during treatment with anti-TNF therapies, especially in high risk regions. The overall rates of AEs did not increase over time, suggesting that CZP treatment may not have a cumulaAliment Pharmacol Ther 2014; 40: 903-916 ª 2014 John Wiley & Sons Ltd

tive dose or duration toxicity over 7 years. Treatments with thiopurine or corticosteroid therapies in contrast have been shown to have an associated risk of cumulative dose and duration toxicity. While long-term treatment with anti-TNF therapy for other indications provides a general understanding of the tolerability, it is important to have primary data from patients with CD because there are some differences in dosing regimens, concomitant medical therapy employed and long-term disease-specific complications compared to other indications. In addition, CD is a life-long condition that requires long-term continual treatment in the majority of patients, so it is critical to be able to understand the implications of chronic continual therapy. The rates of malignancies and skin cancers were low (