A, A GC taper of 3 days or less (n = 17). B, A GC taper longer than 3 days (n = 14). Patient number indicators correspond to the patients listed in Table 4. Patients No. 2 and 32 were removed because they did not receive a taper. X indicates the baseline GC dose.
Lamore RF, Hechenbleikner EM, Ha C, Salvatori R, Harris LH, Marohn MR, Gearhart SL, Efron JE, Wick EC. Perioperative Glucocorticoid Prescribing Habits in Patients With Inflammatory Bowel Disease A Call for Standardization. JAMA Surg. 2014;149(5):459-466. doi:10.1001/jamasurg.2013.5278
Copyright 2014 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.
High-dose glucocorticoids (GCs) are routinely given to surgical patients with a history of GC exposure to prevent perioperative acute adrenal insufficiency, but this practice is not well supported.
To evaluate the variability of perioperative GC dosing among patients with inflammatory bowel disease (IBD) undergoing major abdominal surgery.
Design, Setting, and Participants
This was a retrospective study of 49 patients with IBD undergoing colorectal surgery at a single institution between July 2010 and August 2011. Data on patient comorbidities, intraoperative risk factors, surgical site infections, and 30-day readmission rates were prospectively collected from the National Surgical Quality Improvement Program. Preoperative GC exposure at the time of the index admission and perioperative GC therapy during admission were collected by review of the medical records. Patients were divided into 3 groups at the time of surgery: (1) 1 week or more of prior GC exposure, not receiving maintenance therapy (n = 15); (2) currently receiving budesonide (n = 10); and (3) currently receiving oral prednisone (n = 24).
Main Outcomes and Measures
Perioperative GC exposure was the main outcome. Qualitative comparisons of perioperative exposure stratified by preoperative GC exposure were done. A multivariate logistic regression analysis was performed to determine significant differences in surgical site infection and 30-day readmission rates among patients with and without perioperative GC exposure.
Overall, 38 of 49 patients (78%) received perioperative GCs; intraoperative GCs were administered to 35 of 49 patients (71%), and 33 of 49 patients (67%) received postoperative GCs. Patients received intraoperative and postoperative GCs, respectively, as follows: 8 patients (53%) and 7 (47%) in group 1, 7 (70%) and 3 (30%) in group 2, and 20 (83%) and 23 (96%) in group 3. The median intraoperative GC dose was 100 mg (range, 50-267 mg of hydrocortisone or hydrocortisone equivalent for dexamethasone); the median total postoperative GC dose for the first 5 days after surgery was 485 mg (range, 50-890 mg of hydrocortisone or hydrocortisone equivalent for prednisone). The median duration of postoperative GC administration was 3 days for group 1, 6 days for group 2, and 7 days for group 3. No statistically significant difference in surgical site infection and 30-day readmission rates was detected in the GC exposure vs no-exposure groups.
Conclusions and Relevance
Perioperative GC dosing among patients with IBD undergoing colorectal surgery is highly variable even within a single center. Additional studies are needed to define the risk of postoperative adrenal insufficiency and establish standardized practices for perioperative GC therapy, which may have the benefit of reducing GC overuse.
Systemic glucocorticoids (GCs) are widely prescribed for chronic autoimmune and inflammatory conditions, such as rheumatoid arthritis, systemic lupus erythematosus, asthma, myasthenia gravis, and inflammatory bowel disease (IBD). More than 25 million prescriptions were dispensed for prednisone, an oral GC, in the United States during 2010.1 Moreover, patients with chronic diseases such as IBD frequently undergo surgical procedures for disease control requiring cessation of oral GC therapy at the time of surgery. Because of the concerns about adverse effects of GC withdrawal relating to hypothalamic-pituitary-adrenal axis suppression, these patients are often given high-dose perioperative GC replacement therapy (stress-dose corticosteroids) to prevent secondary adrenal insufficiency during the perioperative or immediate postoperative window. The potential danger of perioperative secondary adrenal insufficiency was first reported2,3 in the 1950s with 2 cases of postoperative mortality from presumed adrenal crisis in GC-dependent surgical patients in whom GCs were withheld at the time of the operation. Subsequently, high-dose perioperative GC therapy administration has become a routine practice and is standard of care in many institutions.
Few cohort studies and randomized clinical trials4- 9 have analyzed the need for supplemental high-dose perioperative GCs in surgical patients, mostly solid-organ transplant recipients, maintained on their baseline immunosuppressive regimen; these studies, along with a 2008 systematic review,10 have suggested that there is no difference in postoperative hemodynamics in patients continuing maintenance GC doses compared with those receiving additional high-dose GCs. Among patients with IBD undergoing major colorectal surgery, there is retrospective evidence11- 13 suggesting that care of patients receiving GCs at the time of the operation can be managed safely with low-dose perioperative GCs and that care of patients with a history of GC exposure alone within 1 year of surgery can be managed safely without perioperative GCs. Furthermore, several publications14- 18 have suggested a variety of perioperative GC supplementation regimens in patients with different GC indications undergoing surgery. In many instances, a history of brief exposure to corticosteroids (ie, 1-2 weeks) was enough to prescribe perioperative high-dose GCs using arbitrary criteria lacking evidence-based studies supporting this practice. On the other hand, systemic GC use is a risk factor for hyperglycemia19 and poor wound healing,20 which may lead to significant complications and morbidity in the perioperative period. Therefore, standardization and guidelines on the appropriate use of perioperative GCs may reduce surgical morbidity. Little is known about current perioperative GC prescribing practices outside of clinical studies. To address this knowledge gap, the aim of this study was to analyze the variability in perioperative GC dosing practices at a single university hospital.
All patients undergoing elective colon or rectal resections at The Johns Hopkins Hospital between July 12, 2010, and August 30, 2011, with an underlying diagnosis of IBD (ulcerative colitis, International Classification of Diseases, Ninth Revision [ICD-9] codes 556.0, 556.1, 556.3, 556.6, 556.8, and 556.9; or Crohn disease, ICD-9 555.0, 555.1, 555.2, and 555.9) were identified in the American College of Surgeons’ National Surgical Quality Improvement Program–targeted procedure colon and rectal modules. Selected colorectal procedures included open and laparoscopic colectomies and proctectomies with inclusion of only major index operations (eg, ileostomy reversals after an index procedure were excluded). Patient comorbidities, intraoperative risk factors, 30-day readmissions, and surgical site infection (SSI) rates were abstracted from hospital electronic medical records and patient follow-up telephone calls by National Surgical Quality Improvement Program clinical reviewers.
Preoperative GC use and perioperative stress-dose GC administration were determined by independent medical record review (R.F.L., E.M.H.). Patients were included if they had been exposed to at least 1 week of oral and/or intravenous GCs for treatment of IBD within the year before surgery. None of the study patients had medical comorbidities requiring immunosuppressive medication. Patients were excluded if they had a preexisting infection at the time of surgical admission. Patients were stratified into 3 groups based on preoperative GC use on admission: (1) 1 week or more of prior GC exposure but not currently receiving therapy, (2) receiving oral budesonide therapy, and (3) receiving oral prednisone therapy. Budesonide is a potent, topically acting GC with minimal untoward systemic adverse effects because of extensive liver metabolism compared with standard GCs, such as prednisone. Perioperative GC exposure or therapy was defined as GCs administered intraoperatively and/or postoperatively. Glucocorticoids given immediately preceding or during the operation were considered intraoperative, and GCs given after the operation were considered postoperative. A postoperative GC taper was defined as consecutively decreasing total daily GC doses until a baseline dose was achieved. The baseline GC dose was defined as the lowest total daily inpatient dose given before hospital discharge. Intraoperative GC prescribing habits were surgeon and/or anesthesiologist dependent; any formal communication or interaction between surgery and anesthesiology teams before the index operation was unknown. Postoperative GCs were prescribed at the discretion of the primary surgical team. All GC doses (except for budesonide) are expressed as milligrams of hydrocortisone or hydrocortisone equivalents (ie, prednisone 1 mg = hydrocortisone 4 mg). In addition, the presence of postoperative hyperglycemia was assessed on the first 7 postoperative days and was defined as a fasting morning glucose level of 180 mg/dL or more (to convert to millimoles per liter, multiply by 0.0555). The Fisher exact test was performed to determine significant differences in SSI and 30-day readmisssion rates among patients with and without perioperative GC exposure. Subsequently, multivariate, stepwise logistic regression analyses were performed to determine significant differences in factors contributing to SSIs and 30-day readmissions, respectively. This study was approved by the institutional review board of The Johns Hopkins University Hospital.
We identified 50 patients with IBD who underwent colorectal resection performed by 4 surgeons (M.R.M., S.L.G., J.E.E., and E.C.W.) with 36 different attending physicians providing anesthesia. One patient died during admission from complications of septic shock and multiple organ system failure and was removed from our analysis. Among the 49 remaining patients, 55% were male (n = 27) and 53% (n = 26) had a diagnosis of Crohn disease (Table 1). The median age was 38 years (range, 18-68 years), with a median body mass index of 25.4 (range, 16.9-47.2 [calculated as weight in kilograms divided by height in meters squared]). Ninety-six percent of the operations (n = 47) were colectomies and 59% of the procedures (n = 29) were laparoscopic. The median length of the operation was 4.1 hours (range, 0.8-12.1 hours). Seventy-eight percent of the 49 patients (n = 38) received perioperative GCs during the index admission. Fifteen patients (31%) had received at least 1 week of GC therapy within 1 year before surgery but were not receiving maintenance therapy on admission (group 1). Ten patients (20%) were receiving budesonide at the time of admission (group 2) and 24 (49%) were receiving prednisone at the time of admission (group 3) (Table 2).
Overall, intraoperative GCs were administered to 71% of the 49 patients (n = 35) and 67% of the patients (n = 33) received postoperative GCs (Table 3). Among the group 1 sample, 53% of the patients (n = 8) and 47% of the patients (n = 7) received intraoperative and postoperative GCs, respectively. In group 2, 70% of the patients (n = 7) were given GCs intraoperatively compared with 30% postoperatively (n = 3). Among group 3 patients, 83% of the individuals (n = 20) received intraoperative GCs and 96% of the patients (n = 23) received postoperative GCs. The median intraoperative GC doses were 75 mg (group 1), 100 mg (group 2), and 100 mg (group 3), with a range of 50 to 267 mg. The median total postoperative GC dose and median duration of postoperative GC therapy were as follows: 400 mg, 3 days (group 1); 650 mg, 6 days (group 2); and 685 mg, 7 days (group 3) (Table 3). The median total GC dose for the first 5 days after surgery was 485 mg (range, 50-890 mg) (data not shown).
There was significant variability in postoperative GC dosing practices, particularly in patients who were receiving prednisone at the time of admission (group 3), as demonstrated in the Figure. The median duration of a postoperative GC taper was 3 days but ranged from 2 to 10 days (interquartile range [IQR], 2 days), reflecting the length of hospital stay. During GC tapering, the median total daily dose per patient was 100 mg (range, 15-400 mg). Thirty-nine percent of the patients (13 of 33) were not given true tapering regimens because the total daily dose was either increased during the taper (8 patients in group 3 and 1 patient in group 1) or the baseline dosing period (3 patients in group 3 and 1 patient in group 2). The median duration of the inpatient baseline GC dose was 3 days (range, 1-18 days; IQR, 3 days) with a median total daily dose of 80 mg (range, 4-80 mg). Twenty-two of 33 patients (67%) who received postoperative GCs were discharged home receiving GCs at a median dose of 80 mg (range, 20-80 mg) (Table 4); 20 of these patients (91%) were receiving prednisone (group 3) at the time of surgery.
Among the 35 patients who received intraoperative GCs, 30 individuals (86%) received postoperative tapers. Among the 5 patients (14%) who did not receive tapers, 1 patient was in group 1 (limited preoperative GC exposure) and 4 were in group 2 (receiving budesonide). There was rarely any medical record documentation of perioperative GC administration; thus, it is uncertain why patients did not receive tapers after having received intraoperative GCs. In addition, there was no indication of clinical concerns for symptoms of secondary adrenal insufficiency or mention of hemodynamic instability requiring fluid resuscitation and/or vasopressor support among these 5 patients.
Eleven of the 49 patients (22%) with IBD were not exposed to perioperative GC therapy during the index admission. Of these 11 patients, 3 individuals (27%) were receiving budesonide, 1 patient (9%) was receiving prednisone, and 7 patients (64%) had a history of prior GC exposure alone at the time of the operation. None of these patients had documentation of postoperative hemodynamic instability requiring fluid resuscitation and/or vasopressor support. Nine of 38 patients (24%) who received perioperative GCs developed postoperative SSIs; 6 of these patients (67%) were in group 3 (receiving prednisone at the time of the operation) with 3 superficial SSIs, 1 deep incisional SSI, and 2 organ space SSIs, and 3 patients (33%) were in group 1 (limited preoperative GC exposure) with 3 superficial SSIs (data not shown). In contrast, only 1 of 11 patients (9%) who did not receive perioperative GC developed an organ space SSI (P = .054). Overall, 3 of 49 patients (6%) had at least 1 laboratory value consistent with hyperglycemia; all of these patients were in group 3 (receiving prednisone at the time of the operation) and received perioperative GCs. One patient with hyperglycemia (33%) developed postoperative pneumonia and 1 patient (33%) had a postoperative organ space SSI. Eight of 49 patients (16%) were readmitted within 30 days. Of 38 patients administered perioperative GCs, 7 individuals (18%) were readmitted compared with 1 of 11 patients (9%) without perioperative GC exposure (P = .07) (Table 1).
Our study confirms that the decision to administer stress-dose GCs to patients with IBD undergoing major abdominal surgery is common. However, we noted highly variable intraoperative and postoperative GC dosing practices. Two patients (6%) with a history of budesonide exposure (group 1) and 4 (11%) receiving prednisone (group 3) received intraoperative hydrocortisone doses of 200 mg or more, well beyond the recommendations suggested by Lewis and colleagues.3 The remaining 29 patients (83%) who received intraoperative GCs were administered doses of 107 mg or less. Among all patients, the total daily dose of GCs during the postoperative taper ranged from 15 mg to 400 mg, and the total postoperative dose ranged from 50 mg to 1490 mg. Nearly 40% (n = 13) of the patients did not receive true postoperative GC tapers. In addition, 5 patients (15%) (group 1, 1 patient; group 2, 1 patient; and group 3, 3 patients) did not receive any GC during a 24- to 48-hour period during their taper for unknown reasons. Five (33%) additional patients either missed a dose or received an extra dose of a GC because of errors such as multiple orders timed closely together or inappropriate timing of order entry (data not shown).
Overall, 38 of 49 patients (78%) received perioperative GCs irrespective of the type of preoperative GC exposure (budesonide vs systemic) as well as the total daily GC dose and duration of use. There were several patients identified in our study cohort who received perioperative GCs who likely did not need them. Nine patients who were either exposed to budesonide in the year before surgery (2 [22%], group 1) or receiving budesonide on admission (7 [78%], group 2) received intraoperative GCs; 4 of these 9 patients (44%) were also given postoperative GCs. Ten study patients (20%) were receiving oral extended-release budesonide preoperatively. Oral budesonide is a topically acting GC with low systemic bioavailability (9%-21%) because of extensive biotransformation in the liver.21 A randomized, double-blind crossover study22 in healthy young adults demonstrated that 5 days of oral budesonide therapy (dosing ranged from 3 to 15 mg/d) resulted in statistically significantly less plasma cortisol suppression compared with prednisolone (20 mg/d). Moreover, budesonide has demonstrated23 fewer GC-related adverse effects compared with systemic GCs, such as prednisolone. In addition to those 9 patients, 7 who were either exposed to prednisone in the year before surgery (n = 1, group 1) or receiving prednisone at the time of admission (n = 6, group 3) were receiving total daily doses of 5 mg or less. Daily doses of 5 mg or less of prednisone for any duration are unlikely to markedly suppress the hypothalamic-pituitary-adrenal axis.24 Using these criteria within our surgical cohort, we conservatively identified 16 of 49 patients (33%) who may not have required perioperative GC therapy based on their GC dosing preoperatively. Arguably, patients with remote exposure to systemic GCs within 1 year of surgery (ie, 1 week of prednisone alone 11 months before surgery) are also unlikely to need high-dose supplemental perioperative GCs. Given that all operations in our surgical cohort were elective, these patients may have benefited from preoperative assessment of adrenocortical function to determine the clinical necessity for perioperative GC supplementation, including a morning serum cortisol level and/or a cosyntropin (corticotropin) stimulation test. A random morning cortisol level is a good first-line test for assessing adrenal insufficiency in otherwise healthy adults with GC exposure. Serum cortisol levels less than 3 µg/dL (to convert to nanomoles per liter, multiply by 27.588) indicate abnormal adrenal function, whereas levels greater than 18 µg/dL rule out adrenal insufficiency.25 Patients with intermediate serum cortisol concentrations (≥3 µg/dL and ≤18 µg/dL) require further dynamic testing of adrenal function, such as the conventional or low-dose corticotropin test or the insulin tolerance test when not contraindicated.
The marked variability of perioperative GC prescribing within our institution has led to the development of an initiative aimed to standardize perioperative GC prescribing habits among our multidisciplinary IBD team and facilitate improved communication between the surgery and anesthesiology teams. A postoperative colorectal surgery GC taper order set was implemented in July 2011, which rapidly tapers GC dosing in patients from the stress dose to their baseline dose during a 30-day period. The taper is intended for patients who have received intraoperative GCs and are clinically stable with no signs of adrenal insufficiency. On days 0 to 1 of the taper, hydrocortisone, 50 mg, is administered intravenously every 8 hours for 3 doses (first dose 8 hours after intraoperative dose); days 1 to 2: hydrocortisone, 25 mg, intravenously every 8 hours for 3 doses; days 2 to 3: hydrocortisone, 25 mg, intravenously every 12 hours for 2 doses; the baseline dose is then initiated. Select patients who are considered at low risk for secondary adrenal insufficiency have preoperative serum cortisol levels determined on the morning of surgery; this includes patients who receive oral budesonide or have a history of GC exposure within 12 months of surgery but who are not receiving GC therapy at the time of admission. Ultimately, patients with normal morning cortisol levels (ie, >18 µg/dL) will not receive high-dose perioperative GC therapy. Furthermore, GC medication errors for 4 of the 5 previously mentioned patients were potentially preventable events that may have been avoided with a standardized perioperative GC order set. A standardized order set also would have avoided the multiple GC orders timed closely together and the inappropriate timing of GC orders that contributed to these medication errors. A computerized order entry system exists at our institution and could be used prospectively to assess use of the perioperative standardized GC order set and corresponding medication error rates. The goal of these interventions is to standardize perioperative GC therapy, reduce dosing errors, and mitigate unnecessary perioperative GC exposure that may be associated with adverse patient outcomes, such as SSIs.
This study has several limitations. It was a retrospective, single-center cohort design in a small number of patients. Given the small sample size, we were not able to establish a statistically significant difference in patient outcomes with and without perioperative GC exposure; however, our data suggest higher postoperative morbidity among the GC-exposed subset. Furthermore, only in-hospital tapers were evaluated because of the unreliability of postdischarge medication records. Thus, these findings likely underestimate the perioperative GC exposure.
Administration of high-dose perioperative GC therapy is a common practice among patients with IBD undergoing major abdominal surgery at a single academic center, but physician prescribing habits are highly variable. Many of the patients who received perioperative stress-dose GCs had a very low clinical risk for postoperative adrenal crisis. These patients may have avoided unnecessary exposure to GCs with a standardized approach to perioperative GC prescribing and/or preoperative assessment of adrenocortical function. Many patients in our study likely had unnecessary perioperative GC exposure; prospective studies are needed to assess the role of perioperative GC dosing in patients with IBD undergoing major colorectal surgery. Furthermore, additional research is required to determine whether standardized perioperative practices, including a preoperative serum cortisol measurement and postoperative GC order set, will result in decreased GC exposure.
Accepted for Publication: June 28, 2013.
Corresponding Author: Elizabeth C. Wick, MD, Department of Surgery, The Johns Hopkins Hospital, 600 N Wolfe St, Blalock Room 658, Baltimore, MD 21287 (firstname.lastname@example.org).
Published Online: March 19, 2014. doi:10.1001/jamasurg.2013.5278.
Author Contributions: Drs Lamore and Hechenbleikner had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Drs Lamore and Hechenbleikner contributed equally.
Study concept and design: Lamore, Hechenbleikner, Salvatori, Harris, Marohn, Efron, Wick.
Acquisition, analysis, or interpretation of data: Lamore, Hechenbleikner, Ha, Salvatori, Marohn, Gearhart, Wick.
Drafting of the manuscript: Lamore, Hechenbleikner, Salvatori, Gearhart, Wick.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Lamore, Hechenbleikner.
Administrative, technical, or material support: Hechenbleikner, Marohn.
Study supervision: Lamore, Ha, Salvatori, Harris, Efron, Wick.
Conflict of Interest Disclosures: None reported.
Additional Contributions: Daniel V. Samarov, PhD, mathematical statistician with the National Institute of Standards and Technology, assisted with statistical analysis and interpretation of data, and Lucy Mitchell, RN, MA, Quality Improvement Specialist and Surgical Clinical Nurse Reviewer, American College of Surgeons’ National Surgical Quality Improvement Program, assisted with acquisition and interpretation of data. None of these individuals received financial compensation for their work.