Association of Preoperative Narcotic Use With Postoperative Complications and Prolonged Length of Hospital Stay in Patients With Crohn Disease

Importance  The use of narcotics among patients with Crohn disease (CD) is endemic.

Objective  To evaluate the association between preoperative use of narcotics and postoperative outcomes in patients with CD.

Design, Setting, and Participants  Patients undergoing abdominal surgery for CD at a tertiary referral center between January 1998 and June 2014 were identified from an institutional prospectively maintained CD database.

Main Outcomes and Measures  Primary end points were overall morbidity, postoperative hospital length of stay, and readmission. Univariate and multivariate analyses were used to assess possible associations between postoperative outcomes and demographic and clinical variables, including preoperative narcotic use.

Results  Of the 1331 patients included, the mean age for patients who had at least 1 pharmacy claim for narcotics within 1 month before surgery was 41.5 years and 41.1 years for patients without a pharmacy claim. Of 1461 abdominal operations for CD, 267 (18.3%) were performed on patients receiving preoperative narcotics. Patients receiving narcotics were more likely to have a current smoking habit (P < .001) with perianal disease (P = .046) and undergoing treatment with biologics (P = .04). Patients with preoperative narcotic use had a longer mean (SD) length of stay (11.2 [8.9] vs 7.7 [5.5]; P < .001) and were more likely to develop postoperative complications (52.8% vs 40.8%; P < .001). Multivariable analysis indicated that preoperative narcotic use was the only independent risk factor associated with both postoperative morbidity (odds ratio = 1.36; 95% CI = 1.02-1.82; P = .04) and prolonged hospital stay (estimate = 2.91; SE = 0.44; P < .001). Subgroup analysis indicated that outpatient narcotic users had increased incidence of adverse postoperative outcomes compared with inpatient-only narcotic users.

Conclusions and Relevance  Preoperative use of narcotics in patients undergoing abdominal surgery for CD is associated with worse postoperative outcomes. Before starting regular narcotic use, patients with CD should be considered for surgical intervention.

Crohn disease (CD) is an immune disorder of the entire gastrointestinal tract that is associated with a high probability of requiring surgery.^1-3 Despite advancements in medical care, the rate of surgery among patients with CD remains high.^1,4-6

One of the most disabling symptoms for patients with CD is chronic abdominal pain that negatively affects quality of life.^7,8 It is estimated that approximately 5% of patients with CD have widespread chronic pain.⁹ A significant proportion of outpatients with CD are also long-term narcotic users,^10-14 as narcotics are increasingly prescribed for chronic pain even for other benign conditions.^15-18 It has been suggested that patients with CD requiring long-term use of analgesics represent a particular CD subgroup that is substantially more difficult to manage.^13,14,19

In addition to decreasing gut transit and intestinal secretion,²⁰ prolonged use of opioids can result in immunosuppression, disarray of the endocrine feedback system, respiratory depression, and mood disorders.^16,21 Among patients with CD, narcotic use has been specifically associated with significant risk of infection,^22,23 increased disease activity, depression, anxiety, and decreased quality of life.^13,14,24 However, there are still scant data on the specific effect of preoperative narcotic use on the outcomes of surgery for abdominal CD, particularly whether preoperative narcotic use is associated with poor postoperative outcomes, such as the presence of disease-related intra-abdominal abscesses or fistulae,^9-11 or with preoperative use of immunosuppressive medications.^3,12,13

The aim of this study was therefore to evaluate possible associations between preoperative narcotic use and adverse postoperative outcomes, including postoperative complications, recovery parameters, and hospital readmissions following abdominal surgery for CD. It was our hypothesis that perioperative use of narcotic analgesic is associated with a higher incidence of postoperative morbidity and increased frequency of unplanned readmission among patients with CD undergoing abdominal surgery.

Box Section Ref ID

Using our institutional review board–approved, prospectively maintained CD database, we identified consecutive patients undergoing abdominal surgery for CD in a tertiary referral center from January 1998 to June 2014. The Institutional Review Board of the Cleveland Clinic approved our study and waived consent because we used a deidentified database. Patients undergoing surgery limited to perianal disease, surgical drainage of abdominopelvic abscess, reconstructive surgery following complications of previous operations for CD, or planned second-stage procedures were excluded. Patients receiving epidural analgesia as part of their perioperative care were also excluded.

A prospectively maintained CD database of patient clinical profiles, including demographics, preoperative disease characteristics, operative details, and postoperative outcomes, was retrospectively reviewed. A standardized data sheet was used to retrieve patient data including disease and complication-specific information included in the database. Demographic information including patient sex, age at time of surgery, body mass index (BMI), disease duration, and smoking history were collected. Disease characteristics included Montreal Classification at the time of surgery, family history of inflammatory bowel disease (IBD), previous CD-related bowel resection(s), coexistent perianal disease, and perioperative blood transfusion. Operative details included the American Society of Anesthesiologists (ASA) score, indication for surgical intervention, intraoperative findings, estimated blood loss (EBL), operative time, and information such as open vs laparoscopic approach and urgent vs elective surgery. Records were also noted for preoperative CD medication use, including steroids, azathioprine/6-mercaptopurine, methotrexate, and biologic agents. In the absence of a widely accepted definition of long-term narcotic use, the study definition of preoperative narcotic use was any electronically documented use of narcotic medications within 1 month before surgery. Urgent surgery was defined as an operation performed on an already-hospitalized patient after failure of inpatient medical treatment.

Primary end points were overall morbidity, postoperative hospital length of stay (LOS), and readmission. Morbidity was counted based on the number of operations after at least 1 postoperative complication. Specific postoperative complications occurring within 30 days after surgery or during the same postoperative hospital stay included surgical site infection (SSI), anastomotic leak, urinary tract infection (UTI), pneumonia, Clostridium difficile infection, small-bowel obstruction (SBO), ileus, thromboembolism, dehydration, acute renal failure, and reoperation. Ileus was defined as the absence of bowel function on postoperative day 5 or the need for nasogastric tube insertion due to abdominal distension, nausea, or vomiting after initiating a liquid diet without evidence of mechanical bowel obstruction.²⁵ Small-bowel obstruction was defined as the presence of obstructive symptoms associated with radiographic findings consistent with mechanical intestinal obstruction.^26,27 Hospital readmissions were counted when occurring within 30 days from the time of hospital discharge. A subgroup analysis was also performed to compare individuals who only received narcotics as inpatients and individuals who had received narcotics as outpatients prior to their hospital admission for elective surgery or leading to surgery for abdominal CD. In this respect, outpatient narcotic use was defined as at least 1 outpatient pharmacy claim occurring within the 30 days preceding surgical intervention.

Statistical analyses were performed using descriptive statistics for demographic data. Descriptive statistics were reported as means and standard deviations for continuous variables and as frequencies for categorical variables. To compare groups for continuous variables, t test (normal distribution) or Mann-Whitney U test for nonparametric distributions of data were used. To compare categorical variables, the χ² test or the Fisher exact test were used. Factors found to be significant on univariate analysis were included in a multivariate logistic regression model to detect independent factors associated with postoperative outcomes. All hypothesis testing was 2-sided, with a P value less than .05 considered statistically significant. All statistical analyses were performed using SAS version 9.3 (SAS Institute).

A total of 1461 abdominal operations were performed on 1331 patients with CD, of whom 267 (18.3%) received narcotics preoperatively. Clinical characteristics of the study population are shown in Table 1. The mean (SD) age at the time of surgery was 41.2 (14.9) years, with a mean (SD) duration of disease since diagnosis of 13.3 (11.6) years. The most common disease presentation was ileocolonic location associated with penetrating disease phenotype, which was identified in approximately 40% of cases. Patients using preoperative narcotics were associated with an increased rate of penetrating disease phenotype, synchronous perianal disease, a history of intestinal resection due to CD, a current smoking habit, and treatment with biological agents. There were not any statistically significant differences between the groups with respect to sex, age, disease duration, BMI, disease location, family history of IBD, preoperative use of steroids or immunomodulators, or perioperative blood transfusion.

Table 1 also shows the perioperative variables. The most common procedure performed was ileocolonic resection, which was performed in almost half of the cases, followed at a distance by total abdominal colectomy (TAC), small-bowel resection, total proctocolectomy (TPC), and segmental colectomy. Most patients underwent surgery with the indication of fistulizing disease, which could be associated at times with disease-related abscess or stricture. Only a minority of the operations were performed urgently. There were no significant differences between groups in the proportion of patients approached with laparoscopic surgery or requiring urgent procedures or stoma creation. In addition, no significant differences were observed between patient groups with respect to indication for surgery, ASA score, specific surgical procedures, intraoperative finding of fistulae or abscess, EBL, or operative time.

The comparative 30-day postoperative outcomes are presented in Table 2. There were 5 overall deaths (0.3%), with a significant association between preoperative narcotic use and postoperative mortality (P = .045). Two of the 3 deceased individuals in the narcotic group developed postoperative abdominal abscesses leading to sepsis and multiple organ failure, while the remaining patient experienced postoperative disseminated intravascular coagulation associated with bowel ischemia. Among patients in the nonnarcotic group, 1 of the 2 postoperative deaths occurred in a patient operated on for bleeding Crohn colitis unresponsive to maximal medical management who continued experiencing coagulopathy, bleeding, and thrombocytopenia postoperatively, leading to multiple organ failure, while the other patient died of postoperative pneumonia leading to respiratory failure. The overall postoperative morbidity rate based on the entire cohort was 43%. Patients who had received preoperative narcotics were associated with increased rates of deep incisional and organ space SSI, UTI, ileus, and thromboembolism compared with patients without preoperative narcotics. While specific rates of reoperation and other specific complications were comparable, patients receiving narcotics before surgery had a significantly longer LOS than those without exposure to narcotics (mean [SD], 11.2 [8.9] days vs 7.7 [5.5] days; P < .001) and also experienced an increased rate of hospital readmission when assessed by univariate analysis (P = .008).

Univariate analysis and multivariate logistic regression analyses were used to identify possible independent risk factors associated with postoperative morbidity, longer LOS, and readmission. The 4 independent risk factors associated with postoperative morbidity were urgent surgery, preoperative use of narcotics, increased EBL, and perioperative blood transfusion (Table 3). Preoperative use of narcotics was also an independent risk factor associated with longer LOS in addition to longer operative time, increased ASA score, and urgent surgery (Table 4). The independent risk factors associated with postoperative hospital readmission were increased EBL, a current smoking habit, preoperative use of steroids, and urgent surgery. However, preoperative use of narcotics was not independently associated with unplanned postoperative readmission (odds ratio = 1.32; 95% CI = 0.89-1.95; P = .17).

Outpatient narcotic users were associated with increased incidence of overall morbidity, readmission, SSI, UTI, thromboembolic events, and mortality compared with both inpatient narcotic users and patients without preoperative exposure to narcotics. However, there were no significant differences between the 3 subgroups in terms of reoperation or other specific complications (Table 5).

To our knowledge, this is the first study specifically examining the effect of preoperative narcotic use on postoperative outcomes after abdominal surgical procedures in patients with CD. We found that preoperative narcotic use was associated with an increased risk of 30-day postoperative morbidity and prolonged hospital stay. Albeit a rare occurrence in the overall study population, even mortality was significantly increased among patients receiving preoperative narcotics. Importantly, multivariable analysis confirmed that preoperative narcotic use was an independent factor associated with increased postoperative morbidity and prolonged hospital stay.

While a number of studies have indicated a high prevalence of narcotic analgesic use among patients with IBD, especially patients with CD,^{10-14,19,24,28,29} few studies have focused on the specific effects of preoperative narcotic use on postoperative outcomes or have indicated that preoperative narcotic use is an independent predictor associated with both continued postoperative narcotic use and increased length of hospital stay.^30,31 Currently, the definition of long-term narcotic use is inconsistent.^10,13,14,32 In our study, we defined preoperative narcotic use as having at least 1 pharmacy claim within 1 month before surgery. Although it might be argued that continuous or frequent intermittent narcotic use for a longer time, such as 3 or 6 months, may better reflect long-term narcotic use, a stricter definition would have substantially decreased our study population. The selection of less strict criteria for the definition of narcotic analgesic use actually emphasizes the significance of our results, if one assumes that at least some patients might be questionably defined as long-term narcotic users and could therefore have diluted the differences between 2 groups truly determined by continued, rather than sporadic, preoperative opioid use. Our finding that the preoperative use of narcotics was a predictor for postoperative morbidity might further support the reported hypothesis that narcotic use per se is an indicator of more severe CD.^11,13,14,24 While preoperative narcotic use was independently associated with increased morbidity and longer LOS, this remains difficult to definitely establish based on our study, where the use of narcotics was also significantly associated with recognized markers of more aggressive disease, such as smoking or penetrating phenotype.

Data from the Crohn’s Therapy, Resource, Evaluation, and Assessment Tool registry studies suggest that narcotic use is a significant predictor of serious infections in patients with IBD,^22,23 and there is separate evidence indicating that narcotic use in patients with IBD is also associated with the incidence of pneumonia.³³ Similarly, our data strongly suggest that preoperative narcotic use is associated with an increase in infectious complications such as SSI and UTI. One of the mechanisms supposed to explain the association between narcotic analgesics and infection is narcotic-induced immunosuppression¹⁶ through opioid receptors expressed by immunomodulating cells.³⁴

Postoperative infections were not the only adverse events associated with preoperative narcotic use. Not surprisingly, patients receiving preoperative narcotic analgesics also experienced a longer LOS. This might depend in part on the increased rate of various postoperative complications associated with preoperative narcotic use, but it also and more directly depends on the increased risk of postoperative ileus,^35,36 which is attributed to µ-opioid receptor binding in the central and enteric nerve system.³⁷ It is recognized that opioid agonists decrease intestinal motility^38,39 and reduce intestinal secretions,⁴⁰ which can then lead to nausea, vomiting, constipation, secondary intestinal pseudo-obstruction, and gastroparesis.⁴¹

One salient finding in the present study that has practical implications is that outpatient use of narcotics was even more markedly associated with the risk of adverse outcomes than the whole group of narcotic users. While it might be argued that inpatient narcotic use is necessary to manage abdominal pain in patients with more severe CD presentations requiring hospital admission, the outpatient use suggests instead that narcotic prescription was deliberately selected instead of an alternative, narcotic-free medical management or an earlier recommendation for surgery.

It is possible that our study overlooks cases of successful medical management inclusive of narcotics by selecting patients who underwent surgery and therefore could not experience successful medical treatment by definition. However, it is widely recognized that it is difficult to stop narcotic analgesics once they are introduced as therapeutic agents.¹⁰ Therefore, the decision in favor of earlier upfront surgery has the potential to prevent the initiation of narcotic use, at least in some patients with chronic abdominal pain and known active CD unresponsive to medical management. For patients already receiving long-term narcotics at the time of the initial surgical evaluation, the use of dedicated care pathways, including the use of laparoscopic surgery when possible, could at least mitigate some of the adverse postoperative sequelae in this patient population.^42,43 In this respect, further studies on the possible benefit of such pathways on postoperative morbidity and LOS are necessary.

The main strength of our study is that we were able to evaluate the association between preoperative narcotic use and postoperative outcomes in a large number of consecutive patients undergoing abdominal surgery for CD. However, we must also recognize several limitations. First, the current study is a retrospective analysis based on a single tertiary care center population that might be associated with increased disease severity and therefore increased risk of concurrent narcotic use, which might be more prevalent than it would be among patients undergoing surgery for CD in general. Second, our study does not include data on other potentially relevant psychotropic medications or the specific doses of preoperative narcotics, so a possible dose-dependent effect of narcotics on postoperative outcomes could not be evaluated. Third, the use of narcotics was based on the medical records of a single institution; therefore, some of the individuals classified as “narcotic-free patients” may have received narcotic analgesics elsewhere, although we made every effort to investigate the medication history from the available medical records. Finally, factors such as preoperative C-reactive protein level, serum albumin concentration, and hematocrit level, which might be potentially associated with postoperative outcomes, were not prospectively collected.

Preoperative narcotic use in patients with CD is associated with an increase in postoperative morbidity and prolonged hospital stay. Before starting regular and continuous narcotic use to manage abdominal pain, patients with CD should be considered for surgical intervention.

Corresponding Author: Luca Stocchi, MD, Desk A 30, Department of Colorectal Surgery, Digestive Disease Institute, Cleveland Clinic, 9500 Euclid Ave, Cleveland, OH 44195 (stocchl@ccf.org).

Accepted for Publication: December 1, 2015.

Published Online: February 24, 2016. doi:10.1001/jamasurg.2015.5558.

Author Contributions: Dr Stocchi had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Li, Stocchi, Remzi.

Acquisition, analysis, or interpretation of data: Li, Stocchi, Cherla, Liu.

Drafting of the manuscript: Li, Stocchi, Cherla.

Critical revision of the manuscript for important intellectual content: Li, Stocchi, Liu, Remzi.

Statistical analysis: Liu.

Administrative, technical, or material support: Li, Cherla, Remzi.

Study supervision: Stocchi.

Conflict of Interest Disclosures: None reported.

Funding/Support: Dr Stocchi is supported by the Story Garschina Endowed Chair. Dr Li is a research fellow at the Department of General Surgery, Jinling Hospital, Medical School of Nanjing University, Nanjing, China.

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.