Efficacy and complications of adalimumab treatment for medically-refractory Crohn’s disease: analysis of nationwide experience in Scotland (2004–2008)


Dr G.-T. Ho, Gastrointestinal Unit, Molecular Medicine Centre, University of Edinburgh, Western General Hospital, Edinburgh, EH4 2XU, UK.
E-mail: gwotzerho@aol.co.uk


Background  Adalimumab is a second generation humanized anti-tumour necrosis factor (TNF) monoclonal antibody with established efficacy in Crohn’s disease (CD).

Aims  To evaluate the efficacy and safety of adalimumab on a nationwide clinical setting.

Methods  We used the Scottish Society of Gastroenterology network to identify and follow up the clinical outcomes of patients with CD treated with adalimumab over a 4-year period (2004–2008).

Results  A total of 98 patients received adalimumab - 100.5 patient follow-up years were recorded (64.3% females; median age at diagnosis of 20.7 years; 88.8% treated with 80/40 mg induction regimen. Eighty eight (89.8%) had previous infliximab with 29 (32.9%) primary nonresponders; 32 (32.6%) were corticosteroid-dependent; 47 (47.9%) were intolerant/resistant to most immunosuppressive therapies (two or more). In all, 60% of patients were in clinical remission at 1-year follow-up, with 30% and 55% requiring dose escalation to weekly therapy at 1-and 2-year follow-up respectively. Overall, 29 (29.6%) patients developed complications with eight nonfatal serious (8.2%) adverse events and 2 (2.0%) case fatalities (sepsis following perforation and disseminated colorectal cancer, respectively).

Conclusions  Adalimumab is efficacious in severe and refractory CD in the clinical setting, although there remain significant therapy- and disease-related risks of serious complications.


Monoclonal antibodies targeting anti-tumour necrosis factor (TNF) are highly effective therapies in Crohn’s disease (and ulcerative colitis).1–4 The use of infliximab (Remicaide; Schering Plough, Kenilworth, NJ, USA), a chimeric monoclonal anti-TNF is established in the current management of IBD. Recent controlled trials have now also demonstrated the efficacy of adalimumab (Humira; Abbott Laboratories, Abbott Park, IL, USA), a fully humanized subcutaneously delivered monoclonal antibody (immunoglobulin G1, IgG1) in Crohn’s disease for induction and maintenance treatment.5–7 Both are currently licensed for the treatment of Crohn’s disease.

Recently, the safety, as well as the efficacy of anti-TNF therapies in the clinical setting, has become the primary focus of many clinicians.8–11 Although infective complications are well-recognized side-effects of anti-TNF use (and other immunosuppressive and corticosteroid therapies) in Crohn’s disease,12 other infrequent but serious toxicities are increasingly reported. Most pertinently, there are now an excess of malignancies implicated in rheumatoid arthritis,13 lung cancers in chronic obstructive pulmonary disease14 and clusters of reported cases of demyelination15 following anti-TNF treatment. The most disconcerting development has been the reports of rare, but uniformly fatal hepatosplenic T-cell lymphomas (HSTCL) in those treated with infliximab and thiopurines, which only came to attention after a very long surveillance period in paediatric Crohn’s disease.16 This has also recently been reported in three IBD patients treated with adalimumab.17 These safety factors and others (immunogenicity for example) have led to many changes in the treatment strategies of anti-TNF in Crohn’s disease even in the very short period since its inception into clinical practice.18 These relatively uncommon safety issues, however, need to be counterbalanced with the known toxicities of the established therapies in Crohn’s disease such as corticosteroids and immunosuppressants. For examples, the Cochrane review reported the number needed to harm (NNH) of only 19 patients with azathioprine use in Crohn’s disease.19 The Mayo Clinic series of 500 patients treated with infliximab showed that only concomitant use of corticosteroids and immunosuppressants was significantly associated with development of serious infective adverse events.8

The availability of alternative anti-TNF treatments has undoubtedly increased the therapeutic armamentarium in IBD; in particular, for the difficult group of patients that has demonstrated loss of efficacy or intolerance in conjunction with infliximab use.20–22 Despite published clinical trials such as GAIN, which assessed the short term outcome of adalimumab therapy in those who have failed infliximab previously5 and CHARM, in moderate-to-severe Crohn’s disease,7 there remains a paucity of safety and efficacy data outside the clinical trial setting. It is also clear that newer anti-TNF agents, such as adalimumab, will be directed in the clinical practice to patients who have failed or lost response to infliximab and other medical-surgical approaches. These patients treated with both infliximab and adalimumab in particular, represent a pre-selected group of patients where clinicians require further guidance in decision-making and risk-counselling to the patients. In this study, we now report on the safety and efficacy of adalimumab in Scotland, UK (5.5 million population) using an established national framework of gastroenterologists (Scottish Society of Gastroenterology, SSG). This is therefore, the first nationwide dataset which provides the ‘real-life’ experience in the use of adalimumab as a second-line anti-TNF therapy in the clinical setting.


Study population

Between 1st October 2007 and 15th February 2008, we (GTH and JS) contacted all gastroenterologists (adult and paediatric), nurse specialists and colorectal surgeons in Scotland by e-mail, letter or over the telephone using the Scottish Society of Gastroenterology (SSG) Network – comprising of 90 active gastroenterologists in 28 hospitals. In this period, 78 patients treated with adalimumab were identified and permissions to review these cases were obtained from the primary physicians. A further audit uptake was conducted between 1st April and 1st May 2008 where further 20 patients were included.

Data collection and analysis

Data were collected retrospectively by case note review on a standardized data collection form. We recorded the disease activity/extent, previous medical therapies and surgical details, development of adverse events and need for further surgery (for active disease). Clinical remission at the point of latest follow-up was defined by the regression of symptoms, in concordance with both patient/physician global assessments (and normalization of inflammatory markers). In-patients who have been previously treated with infliximab, we recorded the indications for switch of anti-TNF therapy. These were: (i) no response – as defined by the lack of response to infliximab up to three infusions with the documentation of clinical signs or symptoms relating to active CD such as stool frequency, abdominal pain, fever, elevated inflammatory markers, recurring drainage or formation of new fistula leading to the cessation of infliximab therapy; (ii) loss of response – if previous history of an initial response to infliximab at least 2 doses of infliximab or more and had lacked improvement or had worsening in at least one subsequent infliximab infusion; (iii) reaction to infliximab – the development of a clinically significant acute or delayed infusion reaction; (iv) Long drug holiday – Previous use of episodic infliximab leading to a significant anti-TNF free period. This was a retrospective case-note review study and Kaplan–Meier survival analyses were used to provide censored data at the time of last data entry, with the latest being 12th August 2008, for the outcome data above. All patients irrespective of previous operations, treatments and disease location were included in this study. Descriptive data were displayed as median values with interquartile ranges. Statistical calculations were performed using Microsoft Excel, GraphPad Prism and Minitab Version 10.


A total of 98 patients were treated with adalimumab between 1st January 2004 and 1st March 2008 (Table 1). There was preponderance of females (65.3%) with a median age at diagnosis of 20.7 years (IQR 14.7–31.5). The median age of treatment with adalimumab was 27.9 years (IQR 20.2–40.8). Eleven (11.2%) patients were aged less than 16 years at initiation. The available median follow-up period was 0.58 years (0.27–1.6), with the maximal duration of therapy at 4.35 years. We have total available patient follow-up years of 914.8 and 100.5 years from diagnosis and the initiation of adalimumab therapy respectively.

Table 1.   Demographics of 98 patients treated with adalimumab in Scotland between 2004 and 2008
n = 98 
Females63 (64.3%)
Median age at diagnosis (years)20.7 (IQR 14.7–31.5)
No (%) <16 years38 (38.8%)
Age at adalimumab therapy (years)27.9 (20.2–40.8)
<16 years11 (11.2%)
Median follow-up period (years)0.58 (0.27–1.6)
≥1 year31 (31.6%)
Disease location
 Ileal15 (15.3%)
 Colonic37 (37.7%)
 Ileocolonic45 (45.9%)
 Oral1 (1.0%)
 Perianal disease25 (25.5%)
 Complex fistulizing disease (Entero-cutaneous, -vesical, -vaginal and –colonic)11 (11.2%)
Previous infliximab therapy88 (89.8%)
Infliximab reaction27 (28.4%)
No response29 (29.6%)
Loss of response16 (16.3%)
Long drug holiday/Others17 (17.9%)
Insufficient data7 (7.4%)
Adalimumab monotherapy (no concomitant immunosuppression)38 (38.8%)

Cohort characteristics: disease severity and previous medical/surgical therapies

Eighty-eight (89.8%) patients had been previously treated with infliximab (median no. of infusions – 3) of which 29 (32.9%) patients had demonstrated no response to infliximab (primary nonresponders). Thirty-six (36.7%) patients had previous surgical resections including 11 (11.2%) with total colectomy and end-ileostomy, 17 (17.3%) with more than two small bowel resections and 7 (7.3%) with multiple perianal abscess and fistulae requiring surgical treatment. Thirty-two (32.6%) patients had prolonged corticosteroid exposure with the inability to be weaned. Forty-seven (47.9%) patients had two or more failed (or were intolerant to) immunosuppressive therapies. Almost all patients (88.8%) had standard 80/40 mg induction regimen, whereas 7 (7.2%) had the higher induction dose of 160/80 mg with respective 2 (2.0%) cases each of 40/40 mg and 24/24 mg induction in some early-onset CD cases. The 24/24 mg induction regimen was used in the two paediatric cases as mentioned earlier based on data from limited paediatric case reports (23). Thirty-seven (38.9%) patients were solely on adalimumab at initiation (e.g. no concurrent immunosuppressive therapy).

Efficacy and dose escalation

Using Kaplan–Meier survival analysis, 17.8%, 42.2% and 56.5% of patients failed to maintain clinical remission at 6-months, 1-year and 2-year follow-up respectively following the initiation of adalimumab. Twenty (20.4%) required further surgery within 1-year of therapy. Censored data also show that 24.4%, 30.7% and 55.0% of patients required dose escalation to 40 mg weekly therapy at 6-months, 1-year and 2-year follow-up respectively. Two (28.6%) patients treated with the higher induction regimen (160/80 mg) required dose escalation over a median follow-up period of 0.56 years (0.24–0.72). We did not observe any differences in the need for dose escalation between patients on adalimumab monotherapy and those on concomitant immunosuppressive therapy (P = 0.54, hazard ratio 0.73).

There were also no significant differences in the rates of clinical remission and need for dose escalation between previous infliximab primary nonresponders (n = 28) and the remainder of the cohort (P = 0.37 and 0.69 respectively). It is noteworthy that censored data show that primary nonresponders have a high failure rate of adalimumab with only 33.4% of patients in clinical remission at 1-year.

Early onset Crohn’s disease

A total of 38 (38.8%) patients had early onset Crohn’s disease (age of diagnosis: 16 years or less) with a median follow-up of 0.76 years (IQR 0.31–1.78). In this sub-group, 37 patients have had infliximab therapy. Eleven (11.2%) patients were commenced on adalimumab before or at 16 years of age with an available median follow-up of 0.41 years (IQR 0.13–1.97). It is of interest that a significant subset in this group (n = 8) were primary nonresponders to infliximab. Three patients were 10 years or younger at commencement of adalimumab. The youngest patient (age of onset – 5.0 years; age of adalimumab therapy – 6.7 years) had colonic disease, was treated with adalimumab for less than 4 weeks. Two further cases (age of onset 5.1 and 6.0 years; age of adalimumab therapy - 9.7 and 10.1 years old) with colonic Crohn’s disease were only treated for 2.5 and 3.0 months respectively on the lower 24/24 mg induction followed by 24 mg maintenance.23

Overall complications

The overall complication rate was 29.6% with 8 (8.1%) patients developing more than one complication (Figure 1). The most common complications were infective and injection pain. In the former group, 6 (6.3%) patients required temporary cessation of therapy and a further case of severe perioral cellulitis which led to permanent discontinuation. In the latter, only one (of six patients) with injection pain required treatment withdrawal. Serious complications necessitating permanent discontinuation of adalimumab occurred in 10 (10.5%) individuals. In this group, all but one patient were females.

Figure 1.

 The complication rates following adalimumab therapy. * Withdrawal – complications leading to permanent discontinuation of therapy.

Fatal complications

There were two deaths in this cohort. A 27-year-old female patient died from sepsis following a perforated ileum (Table 2). This patient had been treated with both infliximab and adalimumab (for 3 months) within the year before her death. She had refractory and severe Crohn’s disease (disease duration 13.6 years) with multiple surgery and attendant problems associated with short gut syndrome. Postmortem showed features of disseminated sepsis with additional evidence of accelerated artherosclerosis, myocarditis and unexpected findings of vacuolation and demyelination affecting the dorsal columns of the upper and mid-thoracic cord. The attending physicians felt that anti-TNF therapy (rather than an adalimumab specific effect) contributed to the death of this patient. In the second case, a 25-year-old female patient had refractory colonic Crohn’s disease (disease duration 4.2 years). She had four infusions of infliximab (2005), 4 years of azathioprine therapy (2003–2007), 3 months of Ciclosporin (2003) and finally, 6 months of adalimumab prior to death (2007). Despite initial improvement, she developed increasing diarrhoeal symptoms and continued to fail. A further limited sigmoidoscopy showed evidence of active disease. She eventually required a laparotomy where a Dukes’ D adenocarcinoma was found in the transverse colon. Further investigations revealed disseminated malignancy and a palliative approach was adopted. There was no family history of cancer and previous colonoscopic examinations did not reveal any polyp or pre-malignant lesions.

Table 2.   Reported complications in-patients treated with adalimumab
  1. F-female, M-male.

Overall complication29 (29.6%)
 Injection pain5 (5.1%)
 Lymphopaenia1 (1.0%)
 Myalgia2 (2.1%)
 Infections requiring short-term treatment withdrawal6 (6.1%)
  2 respiratory tract infection 
  1 tonsillitis 
  1 chlamydial eye infection 
  1 recurrent sore throat with   cervical lymphadenopathy 
 Migraine1 (1.0%)
 Herpes infection1 (1.0%)
 Palpitations (cardiac investigations negative)1 (1.0%)
 Abnormal liver function tests1 (1.0%)
 Unclassified skin rash1 (1.0%)
Complications leading to withdrawal10 (10.2%)
 Lung cancer (69 F) 
 Colonic perforation leading to psoas abscess (22 M) 
 Ileal perforation (27 F) 
 Facial cellulitis (35 F) 
 Central demyelination (34 F) 
 Toxic retinopathy leading to visual loss (22 F) 
 Disseminated colorectal adenocarcinoma (25 F) 
 Paradoxical psoriasis and iritis (20 F) 
 Paradoxical psoriasis (45 F) 
 Injection pain (21 F)  
 Ileal perforation leading to sepsis syndrome (27 F)2 (2.1%)
 Disseminated colorectal adenocarcinoma (25 F)  

Serious complications

There were eight other non-fatal serious complications. We have recently reported in full, the first case of locally-advanced non-small cell lung cancer which resolved spontaneously following cessation of adalimumab therapy in a 70-year-old female smoker.24 We also report a case of extensive and severe central demyelination in a 34-year-old female smoker following 4 months of adalimumab therapy in 2007 (also previously treated with infliximab - 3 infusions in 2004). Following cessation of therapy, high dose corticosteroids and plasmapheresis therapy, this patient improved after 6 months. A 22-year-old female developed concentric loss of peripheral vision (tunnel vision) after 4 months of adalimumab. A clinical diagnosis of toxic retinopathy was made. A similar case has previously been reported in a patient with rheumatoid arthritis following treatment with Etarnecept.25 There were two cases of widespread ‘paradoxical’ psoriasis. The cases of colonic perforation leading to the formation of psoas abscess following colonic perforation and severe facial cellulitis have already been previously described.26, 27 The former was considered secondary to an inadequate response to adalimumab in active Crohn’s disease.


This present nationwide dataset provides direct assessment on efficacy of adalimumab in the ‘real’ clinical setting. Our data confirmed that adalimumab can be efficacious with 60% of patients maintained in clinical remission at 1-year of therapy. This included a significant proportion of primary non-responders to infliximab. The efficacy of adalimumab in the latter group is of significant interest. This finding draws similarity with the rheumatology experience, where alternative anti-TNF is commonly used interchangeably in the failure of a particular anti-TNF treatment. Further studies in primary nonresponse area are clearly relevant. Consistent with our earlier dataset,26 the post-hoc analysis of CHARM and unpublished data from Leuven,28 dose escalation in-patients treated with 80/40 regimen, were necessary in 30% and 55% of patients at 1- and 2-year follow-up respectively.

The more important aspect of this study concerns the safety of anti-TNF therapy by providing a direct assessment of emergent safety signals from a nationwide cohort reflective of the true clinical setting where no pre-selection or referral bias had occurred. Presently, there are little or no safety data on the use of a second-generation anti-TNF therapy (in this case adalimumab) in this subset of patients who have failed most medical and surgical therapies. Taking this into consideration, it is clear that the rate of complication is high with a disquieting case fatality rate of 2.0% (in two young females) and 10.5% serious complication rate of adalimumab over 98.7 patient years of follow-up (2 fatalities/100 patient years). The recent Leuven series of 148 patients treated with adalimumab, who have failed infliximab therapy – with a median duration follow-up of 1.37 years also reported a serious complication rate of 10.8% including two deaths from cardiovascular and lung cancer respectively; and seven cancers in five patients.28 In the North American TREAT registry with 5519 patient-years’ follow-up on 3179 patients treated only with infliximab, the case fatality was 0.48/100 patient years.9 There were seven deaths in our Edinburgh infliximab dataset of patients (212 patients between 1999 and 2007; including 30 with adalimumab therapy) equating to 1.1/100 patient years and 6 (3.5%) cases of malignancies.29 Our data suggest a 2-4 fold excess mortality compared to the TREAT registry and now define a group with a potentially higher risk of developing serious complications and death. However, this must be taken into context with the combined effects of severe refractory disease and significant immunosuppressive exposure. In addition, the inherent difficulties of retrospective data collection and recall bias of this study are emphasized.

Our study immediately poses a few critical questions. First, the nature and extent of the contribution of TNF inhibition to these complications are to be questioned; pertinently, how much of this is a class effect or, specific to adalimumab or infliximab? For examples, the occurrence and resolution of lung cancer and severe demyelination in our cohort followed the initiation and cessation of adalimumab therapy respectively. Both lung cancer and demyelination are high-incident diseases in Scotland as a result of genetic (demyelination) or environmental factors (smoking, and lung cancer) – the present data may also reflect an accelerated course in susceptible individuals after anti-TNF exposure, a class rather than treatment-specific effect. More importantly, is the incidence of rarer complications higher in-patients who have been treated with more than one anti-TNF therapy, purporting a ‘double-hit’ phenomenon? Or is this an epiphenomenon within a high risk population or the effect of a cumulative exposure to two anti-TNF agents?

Collectively, how should clinicians counsel patients of the risks of escalating anti-TNF therapy particularly in this high risk group? In the Mayo clinic dataset, Toruner et al. demonstrated that the respective use of corticosteroids, immunomodulators or infliximab conferred a 3-fold increased risk of developing an opportunistic infection which increased to 15-fold when two or more therapies were used in combination.12 In the similar context, the recent progressive multifocal leukoencephalopathy (PML) cases emphasize the concern of the yet unquantifiable latency effect for these adverse events to unravel in the background of heavy immunosuppressive load.10 This may become increasingly pertinent in the now wider use of both biologics and immunodulators. The above questions may only be answered by the establishment of a dynamic safety registry which is sufficiently robust to address these issues. We believe that the establishment of prospective registering and follow-up of these patients can address this clinical need. The Medicines and Healthcare products Regulatory Agency (MHRA) in the UK has now received 4719 adverse drug reactions to anti-TNF therapies since their introduction into clinical practice, of which 561 were fatal, which (11.9%) underscores this need.

In conclusion, we have demonstrated that adalimumab is efficacious in a high risk nationwide cohort. However, there remain significant known (and unknown) safety concerns in escalating potent medical therapies pertinently, the use of anti-TNF therapy in-patients already at increased risk of disease-related morbidity/mortality. The registering of patients treated with biological therapies should be considered. The British Society for Rheumatology Biologics Register (BSRBR)30 has already demonstrated that this is workable with direct clinical applicability.31 The National Institute of Clinical Excellence (NICE) upon approving the use of anti-TNF agents in rheumatoid arthritis in 2002 included the need for registering the respective use with BSRBR to assess the efficacy and toxicity of these agents in routine use. This recommendation should apply to inflammatory bowel disease as well.


We would like to thank Mr David Armor, Mrs Janice Fennell, Sue Aitken, Elizabeth Lightbody, Joanna Munro and Gaynor Broadley; Dr Alastair MacGilchrist and Peter Mills (Secretary and President of the Scottish Society of Gastroenterology respectively). Declaration of personal and funding interests: G. T. Ho, D. Wilson, I. Arnolt, C. Lees and J. Satsangi have served as speakers for Abbott UK (adalimumab) and Schering Plough (infliximab). J. Satsangi has also received research funding from Schering Plough. Participating Hospitals: Western General Hospital, Edinburgh; Aberdeen Royal Infirmary; Inverness Royal Infirmary; Royal Hospital for Sick Children, Edinburgh; Ninewells Hospital, Dundee; Royal Infirmary of Glasgow; Stobhill Hospital, Glasgow; Western Infirmary and Gartnavel General Hospital, Glasgow; Victoria Infirmary, Glasgow; Yorkhill Children’s Hospital, Glasgow; Monklands General Hospital, Airdrie; Stirling Royal Infirmary; Victoria Hospital, Kirkcaldy; Ayr Hospital, Ayr; Royal Aberdeen Children’s Hospital; St. John’s Hospital, Livingston.