5-Aminosalicylic acid (5-ASA) agents remain the mainstay of therapy for the induction of remission in mild to moderately active UC12
and for the maintenance of remission in UC and possible CD.13
Sulfasalazine, the prototype aminosalicylate, was developed to deliver both an antibacterial agent (sulfapyridine) and an anti-inflammatory agent (5-ASA, mesalamine, or mesalazine); the 5-ASA component of sulfasalazine is primarily responsible for the therapeutic benefit. It is poorly absorbed in the colon and partially absorbed in the small intestine. Several sulfa-free aminosalicylates have been developed in recent years14
to target specific GI sites based on the assumption that the effects of 5-ASA are topical and not systemic. Sulfasalazine and mesalamine have multiple anti-inflammatory effects, including the inhibition of the arachidonic acid pathway along the cyclooxygenase, lipoxygenase, and platelet aggregation factor systems. These drugs are primarily eliminated by the kidneys. Important adverse effects include hypersensitivity reactions, bone marrow suppression, pneumonitis, pancreatitis, and hemolytic anemia. These compounds have a short half-life (6-10 hours) and are extensively metabolized. There is a paucity of clinical data for perioperative use of these medications.
In patients in whom decreased glomerular filtration is more likely (age, >65 years; American Society of Anesthesiologists physical status score, IV or V; revised cardiac risk index score, >2; chronic heart disease), a reasonable approach in the perioperative phase is to discontinue sulfasalazine and mesalamine a day before surgery with resumption 3 days after surgery.15
Patients undergoing proctocolectomy will not require postoperative resumption of these agents.
Glucocorticoid use is common in IBD. The most relevant approach in the perioperative period is to ensure that adequate stress glucocorticoid supplementation is given. In patients with IBD, especially patients with active Crohn disease, glucocorticoids are highly effective in inducing clinical remission. However, the therapeutic role of glucocorticoids in the treatment of IBD is primarily to decrease the intensity of inflammation because they are ineffective in maintaining remission or healing mucosal lesions. Long-term use of glucocorticoids is associated with dependency as well as clinical relapses. In addition, long-term glucocorticoid use is associated with osteopenia and osteoporosis, glucose intolerance and diabetes mellitus, increased intraocular pressure and glaucoma, and severe infections.16
Adequate glucose control must be established in these patients in the perioperative setting because uncontrolled hyperglycemia is associated with worse outcome, including poor wound healing and increased infections.
The increased likelihood of infectious complications was demonstrated in the TREAT (The Crohn's Therapy, Resource, Evaluation, and Assessment Tool) registry, in which glucocorticoid use was an independent factor associated with serious infections (odds ratio [OR], 2.21; 95% confidence interval [CI], 1.46-3.34; P
In a series of 100 patients with IBD who developed opportunistic infections, Toruner et al18
found that glucocorticoid use was significantly associated with the development of opportunistic infections (OR, 3.4; 95% CI, 1.8-6.2). In this series, in multivariate analysis, the risk for opportunistic infection increased substantially with the use of a single immunosuppressant (OR, 2.9; 95% CI, 1.5-5.3) vs a combination of 2 or 3 immunosuppressants (OR, 14.5; 95% CI, 4.9-43.0).
In an epidemiologic study in Olmsted County, Minnesota, glucocorticoid dependence at 1 year was found in 28% of patients with CD and 22% of patients with UC.19
The strategy to minimize adverse effects related to glucocorticoids in patients with IBD is to use the lowest effective dose to induce remission in patients with moderately to severely active CD and acute severe colitis,16,20
along with the early use of other immunosuppressant glucocorticoid-sparing agents and biologic therapy.
In a 1952 article, Fraser et al21
first reported iatrogenic adrenal insufficiency due to preoperative glucocorticoid withdrawal in a surgical patient. The publication of similar findings in a case report the following year led to recommendations for high-dose or “stress-dose” glucocorticoids in the perioperative period.22
Since then, the overall practice has evolved, and currently the doses of glucocorticoid replacement are lower than initially espoused because of concerns about the adverse effects of glucocorticoids, including impaired wound healing, elevated risk of infections, GI bleeding, and hyperglycemia.23
The entire practice of “stress-dose” glucocorticoid replacement, even in its current form, has been questioned by some investigators.22-25
In a recent study, Bruewer et al26
found that patients receiving high-dose and prolonged preoperative systemic glucocorticoid therapy who underwent bowel resection for CD experienced no more postoperative complications than did control patients.
In a series from the University of Tokyo, which analyzed data on all patients with UC from 1963 to 1994, those receiving high-dose glucocorticoids were more likely to undergo colectomy because they were more likely to have a refractory disease and to experience postoperative complications.27
Surgical stress is a potent stimulant of the hypothalamic-pituitary-adrenal (HPA) axis. Stress acts by stimulating the release of corticotropin-releasing hormone and arginine vasopressin that in turn causes adrenocorticotropic hormone (ACTH) release. In surgical patients, the highest ACTH levels are noted during the immediate postoperative recovery period and are likely triggered by trauma and pain. Elevated levels of ACTH are also seen during extubation and reversal of anesthesia and at the time of surgical incision.27-29
Epidural or local anesthesia does not appear to stimulate the HPA axis.30
In addition, plasma ACTH response during surgery is attenuated by opiate drugs.28
Substantial variation exists in the degree of endogenous glucocorticoid release in response to surgical stress. Important factors that may influence this variation include concomitant medication use, antecedent illnesses, and age. In general, patients receiving 5 mg or less of prednisone each day, alternate-day glucocorticoids, or any dose of glucocorticoids for less than 3 weeks are not considered to have a suppressed HPA axis and do not require “stress-dose” glucocorticoids.31-34
In contrast, patients receiving less than 20 mg/d of prednisone (or its equivalent) for more than 3 weeks or with features of Cushing syndrome should be assumed to have a suppressed HPA axis and considered for “stress-dose” glucocorticoid supplementation perioperatively. For patients receiving between 5 and 10 mg of prednisone or its equivalent for more than 3 weeks, a clinical prediction rule cannot be implemented because, in this subset of patients, the HPA axis may or may not be suppressed. Generally speaking, rather than subjecting these patients to a corticotropin stimulation test, which may not reliably predict HPA axis suppression, it may be prudent to provide glucocorticoid supplementation in these patients. However, the administration of glucocorticoids in any patient who has been receiving prednisone doses of more than 5 mg/d for more than 1 week in the 6 to 12 months before surgery appears unnecessary.
The physician performing the preoperative assessment should bear in mind that patients receiving long-term high-dose inhaled or topical glucocorticoids for various conditions may have a suppressed HPA axis and may be candidates for “stress-dose” glucocorticoids. Although a healthy person is estimated to secrete between 20 and 30 mg of cortisol a day,35
the requirement for patients undergoing a surgical procedure, while varying according to the degree of expected stress, rarely exceeds 200 mg of cortisol secretion in 24 hours.22
Currently, expert consensus favors optimizing the glucocorticoid replacement dose to the magnitude of stress posed by the surgery.
Patients receiving glucocorticoid therapy should receive their daily requirement throughout the perioperative period, along with supplementation as outlined in .36,37
Glucocorticoid Management During the Perioperative Period
Purine Analogues (6-mercaptopurine/Azathioprine)
These agents have been widely used as glucocorticoid-sparing agents for maintenance of remission. Both agents are oxidized or methylated in erythrocytes or the liver. The perioperative use of immunomodulators such as purine analogues does not affect surgical outcomes or morbidity. Early complications after proctocolectomy with ileal pouch-anal anastomosis were not found in patients using azathioprine/6-mercaptopurine but were observed in patients receiving high-dose glucocorticoids in a study by Mahadevan et al.38
However, potential concerns related to purine analogues are pancreatitis, leukopenia, hepatitis, and bone marrow suppression.
Azathioprine is known to be antagonistic to neuromuscular blocking agents. Dretchen et al39
described a reversal of neuromuscular blockage produced by d-tubocurarine but an increase of the neuromuscular blockade produced by succinylcholine. The effects of azathioprine on neuromuscular transmission are considered to be secondary to inhibition of phosphodiesterase in the motor nerve terminal. Gramstad40
suggested that the initial dose of neuromuscular blocking drugs in renal transplant patients should be increased in the presence of azathioprine (atracurium by 37%, vecuronium by 20%, and pancuronium by 45%). In this study, the atracurium dose was unaffected by renal function, whereas dose requirements for vecuronium and pancuronium were reduced by 23.2% and 61.5%, respectively, compared with patients not taking azathioprine. These findings suggest a transient antagonism of neuromuscular blockade in the presence of azathioprine.
Aberra et al41
performed a retrospective cohort study in 159 patients with IBD who underwent elective bowel surgery, including 56 patients receiving glucocorticoids alone, 52 receiving 6-mercaptopurine/azathioprine with or without glucocorticoids, and 51 patients receiving neither glucocorticoids nor 6-mercaptopurine/azathioprine. The rate of postoperative infectious complications was not found to be significantly higher in the group receiving glucocorticoids or a combination of glucocorticoids and purine analogues. In an animal study conducted by Brokowski et al,42
the regeneration of a ureter and renal pelvis from a transected and anastomosed strip of the ureteral wall was not associated with any adverse effect on wound healing in dogs receiving prednisone and azathioprine.
In a study by Colombel et al,43
the rate of postoperative complications was not increased in 207 patients with CD who underwent intra-abdominal surgery while receiving glucocorticoids or immunosuppressive therapy with azathioprine, 6-mercaptopurine, methotrexate, or infliximab.
Myrelid et al44
reported an 8% risk of postoperative intra-abdominal septic complications in 343 consecutive patients undergoing surgery for CD. Overall, thiopurine therapy was associated with an increased risk of intra-abdominal septic complications (16% vs 6% without therapy; P
=.044). The use of thiopurines was associated with a 24% risk of septic complications in patients who had known risk factors (both preoperative intra-abdominal sepsis and use of colo-colonic anastomosis), 13% in patients with only 1 of these risk factors, and only 4% in patients with none of these risk factors (P
A potential association between the use of thiopurines and myelotoxicity has been reported.45
It has been recommended, on the basis of weak evidence and physiologic considerations, that thiopurines be stopped on the day of surgery and resumed 3 days afterward because of their renal elimination and potential for toxic metabolite accumulation.46
However, the incidence rate of severe myelotoxicity is less than 1% per patient and year of treatment, and the mortality risk is less than 0.1%47
; therefore, although a theoretical risk for perioperative myelotoxicity exists, it is almost negligible. Therefore, our recommendation is to withhold thiopurines on the day of surgery and, if renal function remains normal, to resume within the first 3 postoperative days when oral medications are resumed.
This potent immunosuppressive agent has been used in patients with glucocorticoid-refractory UC as rescue therapy before colectomy and occasionally in patients with CD. Cyclosporine, a potent inhibitor of T cells that is metabolized in the liver by CYP3A, is primarily excreted in bile; however, 6% of the drug is eliminated unchanged in urine. Major adverse effects include nephrotoxicity, seizures, and opportunistic infections. The mortality rate with opportunistic infections can be as high as 3.5%48
; hence, patients receiving cyclosporine and glucocorticoid therapy should be carefully monitored for any signs of infection. These patients should also receive due consideration for Pneumocystis jiroveci
prophylaxis with trimethoprim-sulfamethoxazole.
Preoperative cyclosporine has not been shown to have any detrimental effects during or after surgery. In a small case series of 25 patients, Pinna-Pintor et al49
found no increased postoperative complications in patients treated with intravenous or oral cyclosporine. These findings were consistent with previous studies.50,51
We recommend careful observation of patients receiving cyclosporine for deterioration in renal function and opportunistic infections. At the same time, cyclosporine levels ought to be carefully monitored. Current clinical data are inadequate to support the discontinuation of this drug before and immediately after surgery.
Methotrexate competitively inhibits the enzyme dihydrofolate reductase, impairing DNA synthesis and therefore cellular replication. Evidence to support its use in UC is minimal.52
Methotrexate is excreted by the kidneys, and patients with renal impairment require dose adjustment. Major adverse effects include thrombocytopenia (up to 10%), pneumonitis, and hepatotoxicity. Perioperative considerations include, but are not limited to, an increase in infectious complications; in the setting of renal impairment, a toxic buildup of its metabolites can lead to bone marrow suppression.53,54
In a 1997 study of a population of patients undergoing elective orthopedic surgery for rheumatoid arthritis, Bridges and Moreland55
found increased perioperative complications in a small number of patients. Multiple subsequent studies in patients undergoing orthopedic surgery have suggested an increased risk of postoperative complications, consisting mainly of infections.56-58
Grennan et al59
published a retrospective study of 388 patients with rheumatoid arthritis who underwent surgery while receiving methotrexate and concluded that continuation of methotrexate does not increase the risk of either infections or surgical complications in patients within one year of elective orthopedic surgery. Most studies have not specifically addressed the effect of methotrexate in the perioperative period on renal function.
Concern has been raised about the potential interaction between nitrous oxide used for anesthesia and methotrexate. A substantial interaction between nitrous oxide–based anesthesia and methotrexate in cancer patients undergoing surgery has been demonstrated by in vivo studies. According to these studies, patients who receive methotrexate during the immediate postoperative period (within 6 hours) after nitrous oxide–based anesthesia often develop severe bone marrow depression and mucositis.60
However, no quantitative data for dose-effect interactions are available regarding the combined toxic effects of methotrexate and nitrous oxide.
Most published trials have included patients receiving low dosages of methotrexate (5-10 mg weekly). Only a paucity of data supports the currently recommended dosages of 15 mg weekly with escalation to 20 to 30 mg weekly depending on clinical response and tolerability.61,62
Existing data do not suggest a significant increase in the risk of perioperative infections or impaired wound healing. Given the lack of data, in patients with a history of previous or severe septic complications, it may be reasonable to discontinue methotrexate 1 week before surgery and resume it no sooner than 1 week after or when the wound has successfully healed. The risks vs the benefits of discontinuation should be discussed with the patient, and the potential for flare of the disease should be weighed against the potential of poor wound healing or infectious complications.63
Treatment for IBD has entered an era of biologic response modifiers. One such agent, infliximab, a chimeric monoclonal antibody targeting TNF-α, is the first drug approved by the US Food and Drug Administration for the treatment of CD. Its role in the treatment of refractory CD is well established,64,65
and current data support its use in patients with moderately to severely active UC who have had an inadequate response to conventional therapy.66
Infliximab has a long half-life of 8.0 to 9.5 days and has its own unique adverse effect profile. Several adverse effects have been reported for TNF-α–blocking agents, including reactivation of tuberculosis; an increased risk of sepsis, pneumonia, and fatal and opportunistic infections (eg, invasive fungal infections, listeriosis, Pneumocystis
infections); reactivation of chronic hepatitis B in carriers; worsened chronic heart failure; optic neuritis; demyelination reactions; bone-marrow toxicity infusion reactions; acute and delayed hypersensitivity reactions; and formation of anti–double-stranded DNA.67-70
Despite its potent immunosuppressive effects, preoperative use of infliximab does not seem to increase postoperative complications in patients with UC or CD. In a cohort of 314 patients with CD, 40 of whom received 1 or more infusions of infliximab before intestinal resection, Marchal et al71
found no increase in postoperative infections or prolongation of hospital stay after infliximab infusion. In a study by Jarnerot et al72
using infliximab as rescue therapy in patients with moderately severe UC, 7 patients who received infliximab underwent colectomy without any increase in the postoperative complication rate. On the basis of their retrospective study of 277 patients with CD who received infliximab within 8 weeks of surgery and 4 weeks after surgery in addition to other immunomodulators, Colombel et al43
concluded that infliximab is safe in the perioperative setting. A 30-day postoperative follow-up showed no increase in septic and nonseptic complications.
In a cohort study by Kunitake et al73
of 413 patients with IBD who underwent abdominal surgery, the rate of postoperative complications was similar in the 100 patients who had received infliximab 12 weeks or less before surgery vs those who had not.
Bordeianou et al74
compared 44 patients with UC and symptoms of unremitting disease who were taking infliximab before TPC or a subtotal colectomy with 127 patients who were not using infliximab. The outcomes in both groups were similar: rate of emergent surgery (4.5% vs 0.4%; P
=.98), rate of subtotal colectomy (19.2% vs 18.0%; P
=.99), or rate of ileoanal J pouch reconstruction (53.8% vs 62.0%; P
=.98). The authors concluded that infliximab contributed no increased surgical morbidity in patients with UC.
Kraemer et al75
found that 16 of 19 patients who received 5 mg/kg of infliximab perioperatively during scheduled anal reconstructive surgery for complicated fistulizing anal CD had a favorable outcome, findings similar to those of other studies in the same population. In the TREAT registry study evaluating 6290 patients, Lichtenstein et al76
found that infliximab was not independently associated with increased risk after adjustment for corticosteroid use and disease severity; however, both corticosteroids and disease severity were associated with adverse outcomes. In a recent study by Gainsbury et al,77
infliximab was not associated with increased risk of short-term postoperative complications after proctocolectomy and ileoanal anastomosis.
A retrospective study of 389 patients with CD who underwent ileocolonic resection at the Cleveland Clinic, 60 of whom received infliximab within 3 months before surgery, found an increased rate of postoperative sepsis, abscess, and readmissions in the patients who received infliximab; the authors of this study suggested that these complications might have been prevented by a diverting stoma.78
Of these patients, those using infliximab had an increased rate of early complications (OR, 3.54; 95% CI, 1.51-8.31; P
=.004) or sepsis (OR, 13.8; 95% CI, 1.8-105.0; P
=.011) and an increased need for a 3-stage procedure (OR, 2.07; 95% CI, 1.18-3.63; P
=.011), leading Mor et al79
to conclude that infliximab use has changed the surgical approach to UC by increasing the number of operations. However, in response to an invited editorial comment, they acknowledged that 3-stage procedures are safe in these patients.
In a multivariate analysis of 301 patients with UC who underwent ileal pouch anal anastomosis, 47 of whom received infliximab preoperatively, Selvasekar et al80
reported that infliximab was the only factor independently associated with infectious complications in this group of patients.
Toruner et al18
described a substantial increase in the risk for opportunistic infection in patients taking a combination of 2 or 3 immunosuppressants (OR, 14.5; 95% CI, 4.9-43) vs those taking a single immunosuppressant (OR, 2.9; 95% CI, 1.5-5.3).
A recent study by Cottone et al81
demonstrated a higher risk of severe infections (11%), neoplasms (3%), and mortality (10%) in patients older than 65 years who received TNF inhibitor therapy than in younger patients or in patients of the same age who did not receive such treatment.
Although new anti-TNFα drugs (eg, adalimumab and certolizumab pegol) are available for the treatment of CD, no data regarding their use in the perioperative setting have been reported. The safety profile of these medications appears to be similar to that of infliximab, especially for fistulizing CD.82-88
Another recently introduced agent that has been approved for the management of moderate to severe CD is natalizumab, a humanized monoclonal antibody against the α4 integrin subunit that inhibits leukocyte adhesion and migration to areas of inflammation. However, safety concerns regarding its association with progressive multifocal leukoencephalopathy have limited its use. Data regarding its use in the perioperative setting are unavailable.
Most evidence suggests that infliximab can be used safely in the perioperative period. Divergent data may reflect the higher burden of comorbidity (concomitant immunosuppressant use, increased severity of disease) in patients with adverse outcomes.
Practices being increasingly recommended to improve outcome include staged surgeries with temporary diverting stomas and the selection of 3-rather than 2-stage ileal pouch anal anastomosis.7,89
In a recent review, Beddy et al5
questioned whether it was justifiable to delay surgery in patients who have recently been administered infliximab or to create a proximal diverting stoma purely to deliver biologic medications. Currently, we do not recommend the discontinuation of immunomodulator therapy with anti-TNF agents in the perioperative setting. However, the clinician should be aware of all possible complications, including serious infections, in surgical patients receiving these agents.
Perioperative medication and Thromboembolic Events in IBD
Patients with IBD have long been known to be at increased risk of thromboembolism. In 1936, Bargen and Barker90
at Mayo Clinic reported that 18 of 1500 patients with UC had evidence of extensive arterial and venous thrombosis. Since then, the association of these complications with IBD has been increasingly recognized. Thromboembolic complications in the cerebral and retinal vasculature,91
the portal vein, and peripheral arteries92
have been reported. In a case series of IBD-related thromboembolism from Mayo Clinic, the activity of disease and extent of colonic involvement in patients with UC were found to be associated with increased risk; however, 87% of patients in this study had another risk factor for venous thrombosis, such as hospitalization, immobilization, malignancy, or recent surgery.93
These risk factors are common in patients who are undergoing surgery, and aggressive antivenous thrombosis prophylaxis should be considered in these patients. However, no guidelines for venous thrombosis prophylaxis specifically in IBD patients have been published. In a study by O'Connor et al,94
the event rate of clinical thrombosis after major abdominal and pelvic surgery for patients with UC was noted to be similar to that in patients without UC undergoing similar surgery. The authors of that study concluded that standard prophylaxis in a patient with UC is acceptable to reduce the occurrence of thrombotic events in the perioperative period.
Inherited risk factors for thrombosis, such as the factor V Leiden mutation, the G20210A mutation in the prothrombin gene, and the homozygous C677T mutation in the methylenetetrahydrofolate reductase gene, have not been attributed to increased thrombosis in patients with IBD.95
However, a thorough investigation of the coagulation profile and genetic testing is advisable in younger IBD patients with a first idiopathic thrombotic event.96
Hyperhomocysteinemia, which is considered a risk factor for arterial as well as venous thrombosis, has been found to be more prevalent in patients with IBD97
but has not been found to be a major contributory factor in the development of venous or arterial thrombosis in patients with IBD.98
Using a large outpatient database (Manitoba Health database), Bernstein et al99
demonstrated a 3-fold increased risk of developing deep venous thrombosis (DVT) or pulmonary embolism (PE) in patients with IBD, incidence rates of 31.4 per 10,000 person-years for DVT and 10.3 per 10,000 person-years for PE in patients with CD, and incidence rates of 30.0 per 10,000 person-years for DVT and 19.8 per 10,000 person-years for PE in patients with UC.99
In a more recent study among 13,756 patients with IBD and 71,672 matched controls, Grainge et al100
found that 139 patients and 165 controls developed venous thromboembolism (VTE); patients with IBD had a higher risk of VTE compared with controls (hazard ratio, 3.4; 95% CI, 2.7-4.3; P
<.001; absolute risk, 2.6 per 1000 person-years).
In the inpatient setting, Nguyen and Sam101
evaluated a large national database of 522,704 discharges of non-IBD patients compared with 73,197 discharges of patients with CD and 43,645 discharges of patients with UC; the risk of VTE was 1.5- to 1.8-fold higher among patients with vs without IBD (P
<.001). In patients with IBD, VTE risk was higher among patients with UC or fistulizing CD.101
The Joint Commission National Quality Core measures specifications include the appropriate use in patients undergoing general surgery of pharmacological VTE prophylaxis, including unfractionated heparin, low-molecular-weight heparin (LMWH), factor Xa inhibitor (fondaparinux), and any pharmacological choice combined with a physical measure (stockings or intermittent pneumatic compression devices). Nonpharmacological prophylaxis is used in patients with a high risk of bleeding.102
No studies have specifically evaluated the potential benefit of VTE prophylaxis in hospitalized or ambulatory patients with IBD. Studies to date do not support an increased bleeding risk with moderate doses of anticoagulant medications in patients with active IBD.103
During the perioperative period, we recommend that patients with IBD receive prophylaxis based on American College of Chest Physicians Evidence-Based Clinical Practice Guidelines (8th Edition).104
However, a recent study by Scarpa et al105
demonstrated that standard prophylactic LMWH may be insufficient for VTE prophylaxis in patients with IBD. In a series of patients undergoing major colorectal surgery from 1999 until 2006, they demonstrated an elevated risk of postoperative DVT in patients with UC colitis (OR, 7.4; 95% CI, 1.4-44.4; P
=.017) despite prophylactic anticoagulation with 4000 IU/d of LMWH. The rate of DVT in patients with UC was higher than in patients with colorectal cancer (P
For patients undergoing open surgery, we recommend prophylaxis with 5000 U of subcutaneous heparin 3 times daily, 40 mg of subcutaneous enoxaparin once daily, or 2.5 to 5.0 mg of subcutaneous fondaparinux daily.