The risk of post-operative complications associated with infliximab therapy for Crohn's disease: a controlled cohort study
Dr P. Rutgeerts, University Hospital Gasthuisberg, Department of Gastroenterology, Herestraat 49, 3000 Leuven, Belgium.
Background : By temporarily suppressing the immune response, the anti-tumour necrosis factor agent, infliximab, may increase the risk of peri-operative complications.
Aim : To test this hypothesis for intestinal resection in a cohort of 313 Crohn's disease patients treated with infliximab. Forty received one or more infusions prior to intestinal resection (31/40 within 12 weeks).
Methods : The post-operative events of these patients were compared with those of a control group (infliximab naive) of 39 patients adjusted for age, gender and surgical procedure. Early (10 days) and late (3 months) major or minor complications were identified.
Results : The incidence of early minor (15.0% vs. 12.8%) and major (12.5% vs. 7.7%) and late minor (2.5% vs. 5.1%) and major (17.5% vs. 12.8%) complications and the mean hospital stay after surgery (10.3 ± 4.0 days vs. 9.9 ± 5.5 days) were similar in both groups. A trend towards an increased early infection rate was found in infliximab pre-treated patients (6 vs. 1; P = 0.10), but more patients in this group received corticosteroids and/or immunosuppressives (29 vs. 16 patients; P < 0.05).
Conclusion : The use of infliximab before intestinal resection does not prolong the hospital stay and does not increase the rate of post-operative complications.
In the last decade, the use of immunosuppressive therapy has gained widespread acceptance in the medical management of Crohn's disease.1 The better understanding of mucosal immunity in Crohn's disease and insights into the effect of gut mucosal cytokines in the inflammatory process have led to the development of new therapies with cytokines and anti-cytokines.2 Tumour necrosis factor-α appears to be a key mediator in inflammatory bowel disease. The chimeric monoclonal antibody to tumour necrosis factor, infliximab, has shown efficacy in the short-term treatment of refractory luminal and fistulizing Crohn's disease.3, 4 Repeated administration every 8 weeks maintains the improvement induced by infliximab in both luminal5, 6 and fistulizing7 Crohn's disease. If this therapy is used for long-term maintenance, it is likely that a number of patients will require surgery after having received infliximab in the recent past. Toxicities related to infliximab therapy have emerged. Infectious complications have been reported, including tuberculosis, other opportunistic infections and sepsis.7, 8 The importance of tumour necrosis factor-α in the host defence against intracellular organisms has been known for some time, and it is not unexpected that the neutralization of tumour necrosis factor-α could predispose to a higher infection rate. By neutralizing tumour necrosis factor with infliximab, and thus suppressing the immune response in the pre-operative setting, the risk of post-operative complications, in particular the infection rate, may increase. The aim of this study was to determine whether infliximab in the pre-operative period increases the post-operative complication rates in patients undergoing bowel resection for Crohn's disease.
We studied retrospectively a cohort of 313 consecutive patients with Crohn's disease who received infliximab (Remicade, Centocor, Malvern, PA, USA) at the specialized infusion unit of the University Hospital Gasthuisberg, Leuven, Belgium, between December 1998 and December 2002. The patients' medical history and previous treatment strategies were collected from medical files, as well as our inflammatory bowel disease files. A complete medical database was available for all patients. The indication for infliximab was intractable luminal Crohn's disease or active fistulizing Crohn's disease not necessitating immediate surgery. Patients were given 5 mg/kg infliximab in a 2-h infusion. Patients with luminal disease received one infusion at baseline, and patients with fistulizing disease received a loading dose at 0, 2 and 6 weeks. Thereafter, patients received different numbers of infusions depending on the treatment indication and the clinical response in an episodic treatment schedule. Maintenance schedules were individually tailored by treating physicians. Patients were monitored during and after the infusion for adverse events. Clinical response and delayed adverse events were assessed during further hospitalization, in follow-up clinic visits and by telephone interviews.
From this cohort, 40 patients were identified who had received one or more infusions prior to intestinal resectional surgery. Resection was performed because of symptomatic stenosis or refractory fistulas and/or abscesses, and rarely because of intractable disease (Table 1). All patients underwent small bowel resection, ileo-colonic resection, left colectomy or abdomino-perineal rectal excision. The peri-operative events in these patients were compared with those of a control group of 39 patients who had never received infliximab and underwent surgery in the same period. Crohn's disease patients for the control group were identified from the daily operation lists used to plan the procedures in the operating theatre during the same time period at our institution. The control group was adjusted for age, gender and surgical procedure. The individual performing the selection of the surgical control patients was unaware of the past medical treatment and post-operative complications. We divided complications into major or minor complications, occurring either within 10 days (early) or 3 months (late) of the operation. Major complications included sepsis, anastomotic leak, peritonitis, local fistula or abscess, wound infection, wound failure, severe anaemia and bulbar ulcer bleeding. Minor complications included haematoma, fever, delayed transit, mild infection and intestinal sub-obstruction. Infection rates were examined during the first 10 days after surgery in both groups (early complications) and included catheter sepsis, wound infection, upper airway infection, infectious diarrhoea and yeast infection.
Table 1. Patient characteristics
|Number of patients||40||39|| |
|Mean age (years)||36.0 (16–73)||38.7 (16–73)||N.S.|
|Duration of Crohn's disease (months)||135 (5–276)||147 (2–480)||N.S.|
|First resection||25 (62.5%)||20 (51.0%)||N.S.|
|Previous resections||15 (37.5%)||19 (49.0%)||N.S.|
| Small bowel resection||1 (2.5%)||6 (15.4%)||N.S.|
| Ileo-colonic resection||11 (27.5%)||16 (41.0%)||N.S.|
| Colonic resection||5 (12.5%)||9 (23.1%)||N.S.|
| Proctectomy||2 (5%)||0 (0%)||N.S.|
| 5-Aminosalicylic acid||12 (30%)||21 (54%)||N.S.|
| Budesonide||2 (5%)||6 (15%)||N.S.|
| Methylprednisolone||10 (25%)||6 (15%)||N.S.|
| Azathioprine||18 (45%)||10 (26%)||N.S.|
| Methotrexate||6 (15%)||1 (3%)||N.S.|
| Corticoids and/or immunomodulator||29 (73%)||16 (41%)||< 0.00017|
| Antibiotics||10 (25%)||11 (28%)||N.S.|
|Indication for surgery|
| Stenosis||20 (50.0%)||24 (61.5%)||N.S.|
| Fistula and/or abscess||17 (42.5%)||14 (36.0%)||N.S.|
| Intractable disease||3 (7.5%)||1 (2.5%)||N.S.|
|Total number of infliximab infusions||4 (1–10)||0|| |
|Time to surgery (weeks)|
| < 4||13 (32.5%)|| || |
| < 8||25 (62.5%)|| || |
| ≤12||31 (78%)|| || |
| > 12||9 (22%)|| || |
Statistical analysis was performed using the two-tailed t-test for quantitative variables. Qualitative variables were compared using the chi-squared test or Fisher's exact test as appropriate. Descriptive statistics included the mean and range. A P value of < 0.05 was considered to be statistically significant.
From the total cohort, 40 patients were identified who received one or more infusions prior to intestinal resection. Ten men and 30 women, with a mean age of 36 years (range, 16–73 years), were included in this study group (Table 1). The mean duration of Crohn's disease was 135 months (range, 5–276 months). Thirty-seven per cent of patients (n = 15) had undergone previous intestinal resectional surgery. Seventy-eight per cent of patients (n = 31) underwent surgery within 12 weeks of infliximab infusion. Eighteen per cent of patients (n = 7) received one infusion and 82% of patients (n = 33) received two or more infusions, with a maximum of 10 infusions. All were on concomitant therapy. Most patients were operated on for complications of Crohn's disease (Table 1).
The post-operative events of these patients were compared with those of a control group of 39 patients who did not receive infliximab prior to surgery. Surgery was performed in the same time period by the same group of surgeons (Table 1). The control group was adjusted for age, gender and surgical procedure. These data are also given in Table 1.
Patients in both groups were closely monitored in the post-operative period and vital signs (heart rate, blood pressure, temperature), clinical evolution (abdominal status and wound healing) and biochemical parameters were checked every day until discharge. Further follow-up was assessed by follow-up clinic visits and re-hospitalizations.
The incidence of early minor complications was 15.0% (n = 6) in the infliximab group vs. 12.8% (n = 5) in the control group (Table 2). Early major complications occurred in 12.5% (n = 5) in the infliximab group and in 7.7% (n = 3) in the control group (Table 2). Late minor complications occurred in 2.5% (n = 1) in the infliximab group and in 5.1% (n = 2) in the control group (Table 2). Late major complications occurred in 17.5% (n = 7) in the infliximab group vs. 12.8% (n = 5) in the control group (Table 2). The rates of complications were not significantly different between the two groups. The mean hospital stay after surgery was comparable for both groups (Table 2), i.e. 10.3 ± 4.0 days for the infliximab group vs. 9.9 ± 5.5 days for the control group (N.S.). In the study group, two patients were hospitalized in the intensive care unit for haemodynamic problems caused by catheter sepsis for 1 and 2 days, respectively. In the control group, only one patient was hospitalized for 12 days in an intensive care unit for post-operative suture leak and faecal peritonitis. There was no mortality.
Table 2. Outcome after surgery
| Small bowel resection||6 (15.0%)||6 (15.5%)||N.S.|
| Ileo-colonic resection||17 (42.5%)||17 (43.5%)||N.S.|
| Colonic resection||11 (27.5%)||11 (28.2%)||N.S.|
| Proctectomy||6 (15.0%)||(12.8%)||N.S.|
|Mean duration of hospital stay ± s.d. (days)||10.3 ± 4.0||9.9 ± 5.5||N.S.|
| Early (within 10 days)|| 6 (15.0%)||5 (12.8%)||N.S.|
|Yeast infection: 1*||Haematoma: 3|| |
|Infectious diarrhoea: 2*||Fever: 1|| |
|Fever: 2||Peripheral arthritis: 1|| |
|Upper airway infection: 1*|| || |
|Prolonged ileus: 1|| || |
| Late (within 3 months)||1 (2.5%)||2 (5.1%)||N.S.|
|Yeast infection: 1*||Adhesion + sub-obstruction: 1|| |
|fissure: 1|| |
| Early (within 10 days)||5 (12.5%)|| 3 (7.7%)||N.S.|
|Catheter sepsis: 2*||Anastomotic leak: 2|| |
|Anaemia + transfusion: 1||Faecal peritonitis: 1|| |
|Wound infection: 1*||Candida sepsis: 1*|| |
|Wound failure: 1||Anaemia + transfusion: 1|| |
| Late (within 3 months)||7 (17.5%)||5 (12.8%)||N.S.|
|Anaemia + transfusion: 1||Wound failure: 1|| |
|Idiopathic peritonitis: 1||Bulbar ulcer bleeding: 1|| |
|Local abscess recurrence: 2||Suture leakage: 1|| |
|Persistent fistula: 1||Non-local abscess: 1|| |
|Fistula recurrence: 2||Persistent cutaneous fistula: 1|| |
When the infection rates during the first 10 days after surgery were compared in both groups (early complications), there was a trend towards an increased infection rate in infliximab pre-treated patients (P = 0.10). In this group, there were six patients with eight infectious complications (two patients suffered two different infections). The infections in these patients included catheter sepsis (n = 2), local wound infection (n = 1), upper airway infection (n = 1), infectious diarrhoea (Clostridium difficile, n = 1; toxin-associated diarrhoea, n = 1) and oro-pharyngeal candidiasis (n = 2). Two patients with mildly elevated temperature during the first two post-operative days were not included. In the control group, one patient suffered from Candida sepsis. One patient with elevated temperature on the first post-operative day was not included. In conclusion, there were eight infectious events in the study group vs. one event in the control group (P = 0.03). The number of patients receiving corticosteroids and/or immunosuppressives (azathioprine, ciclosporin, methotrexate) at surgery was significantly higher in the infliximab group (29 patients) than in the control group (16 patients) (P < 0.0002).
Infliximab has been shown to be an effective treatment for patients with Crohn's disease, but there is limited information available on the peri-operative complication rate when this medication is given in a pre-operative setting. In this study, we found no increase in complication rates in the post-operative period in patients with Crohn's disease who received infliximab prior to surgery. The infliximab group was compared with a control group adjusted for age, gender and surgical procedure. There were no significant differences between the two groups when early minor or major and late minor or major complications were compared. There was a slight, but not significant, increase in the proportion of patients suffering from post-operative infections in the infliximab group. The number of patients receiving corticosteroids and/or immunosuppressives (azathioprine or methotrexate) at surgery, however, was significantly higher in the infliximab group. The mean hospital stay after surgery was comparable in the two groups. Although we cannot exclude the possibility that the absence of excess post-operative complications in patients treated with infliximab in the pre-operative setting is caused by a type II error, due to the limited number of patients in our cohort, we believe that the patient group is sufficiently large to detect clinically relevant differences.
Few reports have studied the use of infliximab in the pre-operative period and the post-operative outcome. In a study by Brzezinski et al., reported in abstract form, the rate of post-operative complications was investigated in 35 Crohn's disease patients who received infliximab in the peri-operative period.9 Twenty-two patients received infliximab within 2 months prior to surgery and 13 patients received the infusion 1 month after surgery. These patients were compared with two control groups matched for age, gender, concomitant use of immunosuppression, surgical procedure and surgeon over the same period of time. There were no differences in the length of stay or complication rates when patients who received infliximab before or after surgery were compared with controls. The authors concluded that the use of infliximab in the peri-operative period is safe and not associated with an increased risk of complications.
In a recent retrospective study, Colombel et al. investigated the rate of complications after abdominal surgery for Crohn's disease associated with perioperative infliximab and/or immunomodulator therapy.10 They searched the medical notes of 270 patients who underwent abdominal surgery between October 1998 and December 2001 at the Mayo Clinic, Rochester, MN, USA. They did not find an increased rate of septic complications or overall complications associated with peri-operative steroid use and dose, immunomodulator therapy or infliximab use in comparison with patients who did not receive these medications. In the study by Colombel et al., 52 patients received infliximab within 8 weeks before or 1 week after surgery.
In contrast, Aberra et al. reported a clearly increased risk of infectious complications after elective bowel surgery for inflammatory bowel disease in patients taking corticosteroids and/or 6-mercaptopurine/azathioprine, when compared with patients not taking these medications.11 Patients receiving corticosteroids had adjusted odds ratios for any and major infectious complications of 3.69 [95% confidence interval (CI), 1.24–10.97] and 5.54 (95% CI, 1.12–27.26), respectively. The adjusted odds ratios for patients receiving 6-mercaptopurine/azathioprine for any and major infectious complications were 1.68 (95% CI, 0.65–4.27) and 1.20 (95% CI, 0.37–3.94), respectively.
In the present cohort, we found a slightly increased rate of infectious complications after resections for Crohn's disease in patients treated with infliximab prior to operation. These patients were, however, more exposed to glucocorticosteroids and/or immunosuppression, which may account for this slight excess.
Our data suggest that patients treated with infliximab can undergo resectional surgery safely without an increased risk of complications or prolonged hospital stay. Whether infliximab therapy facilitates surgery in Crohn's disease needs to be investigated further.
This study was supported by an educational grant from Centocor, Malvern, PA, USA.