Ingraham AM, Cohen ME, Bilimoria KY, Ko CY, Hall BL, Russell TR, Nathens AB. Effect of Delay to Operation on Outcomes in Adults With Acute Appendicitis. Arch Surg. 2010;145(9):886-892. doi:10.1001/archsurg.2010.184
Copyright 2010 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2010
To examine the effect of delay from surgical admission to induction of anesthesia on outcomes after appendectomy for acute appendicitis in adults.
Retrospective cohort study with the principal exposure being time to operation. Regression models yielded probabilities of outcomes adjusted for patient and operative risk factors.
Data were submitted to the American College of Surgeons National Surgical Quality Improvement Program database from January 1, 2005, through December 31, 2008.
Patients with acute appendicitis who underwent an appendectomy.
Main Outcome Measures
Thirty-day overall morbidity and serious morbidity/mortality.
Of 32 782 patients, 24 647 (75.2%) underwent operations within 6 hours of surgical admission, 4934 (15.1%) underwent operations more than 6 through 12 hours, and 3201 (9.8%) underwent operations more than 12 hours after surgical admission. Differences in operative duration (51, 50, and 55 minutes, respectively; P < .001) were statistically significant but not clinically meaningful. The length of postoperative stay (2.2 days for the >12-hour group vs 1.8 days for the remaining groups; P < .001) was statistically significant but not clinically meaningful. No significant differences were found in adjusted overall morbidity (5.5%, 5.4%, and 6.1%, respectively; P = .33) or serious morbidity/mortality (3.0%, 3.6%, and 3.0%, respectively; P = .17). Duration from surgical admission to induction of anesthesia was not predictive in regression models for overall morbidity or serious morbidity/mortality.
In this retrospective study, delay of appendectomy for acute appendicitis in adults does not appear to adversely affect 30-day outcomes. This information can guide the use of potentially limited operative and professional resources allocated for emergency care.
Appendectomy is the most common emergent surgical procedure performed worldwide, with appendicitis accounting for approximately 1 million hospital days annually.1 The pathogenesis of appendicitis stems from mucous production, bacterial overgrowth, and luminal obstruction, causing increased appendiceal wall tension and leading to eventual necrosis and perforation. Increased time from onset of symptoms to operative intervention is associated with more advanced disease.2- 6
Recent developments in imaging and antibiosis have afforded improved preoperative assessment and treatment, allowing for nonoperative management of abscesses and phlegmons and potentially limiting the need for immediate operative intervention to halt disease progression.7 Reports8,9 in the pediatric literature suggest that delaying surgery while providing fluids and antibiotics is not associated with untoward patient outcomes. Studies5,10,11 investigating the effect of surgical delays on subsequent outcomes in adults, however, are limited by small sample sizes and single-institution data. Randomized controlled trials have suggested that antibiotic therapy without surgical intervention appears to be a safe, first-line treatment of acute appendicitis in carefully selected men.12,13 Failure of this conservative strategy may lead to delays in surgical intervention. Given these considerations, the objective of this study was to assess the effect of delays from surgical admission to induction of anesthesia on 30-day outcomes in adults with acute appendicitis.
The American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) provides validated, risk-adjusted surgical morbidity and mortality outcomes, affording private-sector hospitals the ability to conduct in-depth, masked quality comparisons with other participants.14 The ACS NSQIP samples the first 40 procedures performed within consecutive 8-day cycles from general surgery, vascular surgery, and specific subspecialty procedures. Specially trained surgical clinical reviewers abstract medical records and personally communicate with patients to obtain comprehensive clinical data regarding patient demographics, preoperative risk factors and laboratory values, operative information, and perioperative and postoperative outcomes within 30 days of the index operation. On-site audit programs standardize data collection and ensure data consistency and reliability.15- 17
Patients 16 years or older who had undergone an appendectomy based on primary Current Procedural Terminology (CPT)18 codes (laparoscopic: 44970; open: 44950 and 44960) and had a postoperative diagnosis of acute appendicitis based on International Classification of Diseases, Ninth Revision (ICD-9) diagnosis codes (ICD-9 codes 540, 540.0, 540.1, and 540.9) were identified from the ACS NSQIP database from January 1, 2005, through December 31, 2008. Because the ACS NSQIP does not abstract information regarding pathology results and the population of interest is patients with acute appendicitis, patients must have both undergone an appendectomy and had a postoperative diagnosis of acute appendicitis to be included in this study. Patients who underwent an incision and drainage of an appendiceal abscess (CPT codes 44900 and 44901) or an appendectomy concurrent with another major procedure (CPT code 44955) were not included. Patients were classified as having simple vs complicated appendicitis according to the presence (ICD-9 codes 540.0 and 540.1) or absence (ICD-9 codes 540 and 540.9) of generalized peritonitis or abscess.
Duration from surgical admission to induction of anesthesia was studied as a continuous variable and also when restructured into a categorical variable. Data from patients were divided into 3 categories based on time from surgical admission to induction of anesthesia: 6 hours or less, more than 6 hours through 12 hours, and more than 12 hours. Patients with an anesthesia start time before the time of admission (ie, who were transferred from the emergency department directly to the operating room without first being admitted) were categorized with the 6-hour-or-less group.
The primary outcomes of interest were 30-day overall morbidity and 30-day serious morbidity/mortality. Overall morbidity was defined as having documentation of a serious morbidity or at least 1 of the following ACS NSQIP postoperative complications: superficial surgical site infection, deep surgical site infection, pneumonia, unplanned intubation (without preoperative ventilator dependence), peripheral neurologic deficit, urinary tract infection, and deep vein thrombosis. Serious morbidity/mortality was defined as having documentation of mortality or at least 1 of the following ACS NSQIP postoperative complications: organ space surgical site infection, wound dehiscence, neurologic event (cerebrovascular accident or coma lasting >24 hours), cardiac arrest, myocardial infarction, bleeding requiring transfusion, pulmonary embolism, ventilator dependence of more than 48 hours, progressive or acute renal insufficiency, and sepsis or septic shock. Patients could experience more than 1 of the listed complications. Patients were excluded from having the following complications if the condition was documented preoperatively: wound infection, pneumonia, ventilator dependence, renal failure, stroke, and coma. Surgical clinical reviewers examine medical records, attempt to contact patients a minimum of 3 times via telephone or mail, and query the Social Security Death Index and the National Obituary Archives to determine mortality.
Variables for 30-day overall morbidity and serious morbidity/mortality risk adjustment included disease severity, laparoscopic vs open approach, age, sex, race, body mass index (calculated as weight in kilograms divided by height in meters squared), American Society of Anesthesiologists class, and preoperative functional status (as defined by the ACS NSQIP). We considered the possibility that selection of a laparoscopic vs open approach might relate to timing of the operation, with a bias toward open procedures if performed later (owing to perceived difficulty). Thus, the models were run with and without operative approach. We also considered the possibility that the severity of appendicitis might be a result of delay (rather than a baseline covariate). Thus, the models were also run with and without this covariate. In both instances, the results were similar. For clarity, the results are presented with operative approach and postoperative diagnosis in the model. Comorbidities included diabetes mellitus, renal disease, dyspnea, ascites, chronic obstructive pulmonary disease, current pneumonia, ventilator dependence, long-term steroid use, bleeding disorders, heart failure, hypertension, coronary artery disease, peripheral vascular disease, disseminated cancer, weight loss, current chemotherapy or radiotherapy, neurologic deficit, preoperative transfusion, and preoperative systemic inflammatory response syndrome, sepsis, or septic shock. Behavioral risk factors included alcohol use and smoking. All risk factors were converted into discrete categories.
Eleven preoperative laboratory variables were also included (sodium, creatinine, albumin, bilirubin, aspartate aminotransferase, alkaline phosphatase, white blood cell count, hematocrit, platelet count, partial thromboplastin time, and international normalized ratio). Categories of laboratory values were constructed using the ACS NSQIP definitions of normal and abnormal19; missing data comprised a third category.
Demographic characteristics, preoperative risk factors, and rates for overall morbidity and serious morbidity/mortality of patients who underwent appendectomy at the previously described categorical time intervals were compared using χ2 tests for categorical variables and analysis of variance for continuous variables with significance set at P ≤ .05. Forward stepwise multivariable logistic regression models, predicting overall morbidity and serious morbidity/mortality, were constructed. For each outcome, variables selected for inclusion in the logistic regression were included in a random intercepts, fixed slopes hierarchical model using SAS PROC GLIMMIX (SAS Institute Inc, Cary, North Carolina). Time to operation was evaluated as a categorical and continuous variable. If time from surgical admission to induction of anesthesia was not selected in stepwise logistic regression, this variable was forced into logistic and hierarchical models. All analyses were performed using SAS statistical software, version 9.2 (SAS Institute Inc).
During the study period, 32 782 patients underwent appendectomy for acute appendicitis. Of the 32 782 patients, 24 647 (75.2%) underwent operations within 6 hours of surgical admission, 4934 (15.1%) underwent operations more than 6 through 12 hours after surgical admission, and 3201 (9.8%) underwent operations more than 12 hours after surgical admission. Patients on average were 38.05 years old (SD, 16.21 years) and predominantly male (17 672 [53.9%]) (Table 1). Males had shorter durations from surgical admission to induction of anesthesia than females. White patients had shorter durations from surgical admission to induction of anesthesia than black patients. Comorbidities, including but not limited to coronary disease, hypertension, and chronic obstructive pulmonary disease, were more commonly present among those with greater times from surgical admission to induction of anesthesia.
Of the 32 782 patients, 27 334 (83.4%) had simple appendicitis and 5448 (16.6%) had complicated appendicitis (Table 2). Of the 32 782 patients, 25 021 (76.3%) underwent a laparoscopic procedure and 7761 (23.7%) underwent an open procedure. The mean (SD) time to induction for the entire group was 4.95 (42.12) hours. The mean (SD) time between surgical admission and induction of anesthesia was 1.51 (1.72) hours for the 6-hours-or-less group, 8.45 (1.69) hours for the more-than-6- through 12-hour group, and 26.07 (132.62) hours for the greater than 12-hour group. The duration of operation (55 minutes for the >12-hour group, 51 minutes for the ≤6-hour group, and 50 minutes for the >6- through 12-hour group; P < .001) was statistically, but not clinically, significantly different among the 3 groups.
Overall length of stay (from surgical admission to discharge) was statistically significantly different among the 3 groups (1.8 days for the ≤6-hour group, 2.0 days for the >6- through 12-hour group, and 3.1 days for the >12-hour group; P < .001). Length of postoperative stay (2.2 days for the >12-hour group vs 1.8 days for the remaining 2 groups; P < .001) was also statistically significantly different among the 3 groups but not clinically significant. Unadjusted overall morbidity rates were highest in the greater-than-12-hour group (6.6%, n = 210), although there was not a statistically significant linear trend across the time intervals (Cochran-Armitage trend test P = .21) (Table 3). Unadjusted serious morbidity/mortality rates were highest in the more-than-6- through 12-hour group (120 [2.4%]); again, the linear trend was not significant (Cochran-Armitage trend test P = .08). No significant differences were found in the adjusted rates of overall morbidity (P = .33) and serious morbidity/mortality (P = .17) across the 3 time intervals. In addition, when the operative approach (laparoscopic or open) and postoperative diagnosis were not considered in logistic models, the mean estimated risks of overall morbidity (5.6%, 5.6%, and 6.0%, respectively; P < .001) and serious morbidity/mortality (3.0%, 3.1%, and 3.5%; P < .001) were statistically but not clinically significant across the 3 groups.
Duration from surgical admission to operation, as either a categorical or continuous variable, was not selected for inclusion in the logistic regression models for either overall morbidity (C statistic = 0.71; Hosmer-Lemeshow P = .11) or serious morbidity/mortality (C statistic = 0.78; Hosmer-Lemeshow P = .01) and thus was not an important predictor of either of these outcomes. Complicated appendicitis; higher American Society of Anesthesiologists class; systemic inflammatory response syndrome, sepsis, or septic shock; abnormal sodium levels; abnormal creatinine levels; poor nutritional status (albumin <3.5 g/dL [to convert to grams per liter, multiply by 10]); black race; functional dependence; smoking; and dyspnea were associated with an increased risk of having either any morbidity or serious morbidity/mortality (data not shown). The laparoscopic approach was associated with a lower risk (odds ratio, 0.61; 95% confidence interval, 0.54-0.68) of overall morbidity after appendectomy but was not selected for inclusion in the serious morbidity/mortality model.
In this study, we examined the association between 30-day overall morbidity and serious morbidity/mortality after appendectomy for acute appendicitis and the duration between surgical admission and induction of anesthesia. Patients who had an appendectomy at a time further from surgical admission were older, were more functionally dependent, and had a greater likelihood of having comorbidities. Blacks were more likely to undergo an appendectomy at a later time interval, which may indicate a disparity in prompt access to emergency surgical care for this population. In our study, the statistically significant increases in operative times (approximately 4 minutes longer) for the more-than-12-hour group do not represent a clinically significant burden. We found no significant differences in 30-day adjusted outcomes for patients undergoing surgery at various time intervals from surgical admission to anesthesia start time (≤6 hours, >6 through 12 hours, and >12 hours). The negative influence of disease severity on outcomes and the tendency for disease severity to progress over time increase the probability that a significant relationship between surgical delay and adverse events would be detected. That the results did not change when disease severity was excluded from the models suggests that there truly is no relationship between time from surgical admission and negative outcomes after appendectomy in this retrospective cohort study.
Our findings are consistent with several other studies in the adult and pediatric literature that have found no increased rates of complications among patients who had a delay to appendectomy. Abou-Nukta et al10 demonstrated that adult patients with pathologically confirmed acute appendicitis (n = 380) who underwent appendectomy between 12 and 24 hours compared with less than 12 hours after emergency department admission at a large teaching community hospital did not have statistically significant differences in the length of stay, operative time, or rate of complications. Omundsen and Dennett11 found no difference in the complication rate or time to discharge among patients with pathologically confirmed appendicitis (n = 345) undergoing appendectomy within 12 hours of admission vs 12 through 24 hours but observed higher rates of complications and increased postoperative lengths of stay among those having surgery after 24 hours of admission compared with less than 24 hours after admission. The association of advanced disease and delay to operative intervention with increased hospital length of stay and incidence of postoperative complications was documented at a tertiary urban care hospital in patients with all stages of acute appendicitis.5 Yardeni et al9 demonstrated no significant increase in the rate of complications or perforation rate in 126 pediatric patients with acute, pathologically confirmed appendicitis, which was presumed to be nonperforated on preoperative assessment, when appendectomies were performed within 6, 6 through 12, or greater than 12 hours of admission. Surana et al8 found no statistically significant difference in the rate of complications experienced by children (n = 695) who underwent appendectomy within 6 hours vs 6 through 18 hours after admission.
The results of this study should be considered in light of several limitations. Although delay of appendectomy for acute appendicitis in adults after surgical admission was not observed to adversely affect 30-day outcomes, this parity may have been influenced by differences in patient or organizational factors or by clinical interventions about which we have limited or no data. Thus, the absence of an effect owing to delay must be interpreted in the entirety of the retrospective clinical context.
Specific limitations also need to be considered. First, identification of patients with acute appendicitis was conducted through the selection of specific CPT operative codes in combination with postoperative ICD-9 diagnostic codes; computed tomographic scan imaging or pathology reports are not available in the ACS NSQIP database. Second, the ACS NSQIP does not collect information about patients' symptoms, insurance status, emergency department presentation date or time, or reasons for potential delay to the operating room. Thus, we could not control for any delay owing to a patient deferring initial presentation to the hospital despite having symptoms. The delay unaccounted for by the emergency department evaluation is likely to have a minimal effect, especially because the duration of time prior to a patient's presentation contributes more significantly to disease severity than the delay from hospital presentation to operation.2- 6 Third, the ACS NSQIP does not collect data on timing or appropriateness of antibiotics or intravenous fluids administered before the patient undergoes the appendectomy, which represents standard of care and if not received could result in elevated complication rates. Finally, there might be confounding by an indication that cannot be addressed using our approach. Specifically, the timing of the operation might be made based on unmeasured factors that the surgeon or team identifies when assessing the patient. Thus, earlier operations on patients of concern (or delaying operations on patients without concern) might negate any observable differences in outcome associated with delays.
Because of the growing issues surrounding access to emergency care and specialist coverage,20- 22 care for emergency general surgery patients is increasingly the responsibility of acute care surgeons and specialized services, which cover the specialties of trauma, emergency general surgery, and critical care.23 As the elderly population continues to increase, the medical needs of patients presenting for emergency general surgical care will become increasingly complex and will demand additional resources and attention. Because of potentially limited physical and professional staffing resources, an acute care surgeon may need to delay the operation of less critically ill patients to appropriately care for those requiring immediate attention. Our research demonstrates that the frequent, though previously minimally researched, practice of nonimmediate operative treatment of adult patients with acute appendicitis does not appear to significantly affect patient outcomes.
In conclusion, delay of appendectomy after surgical admission does not appear to adversely affect 30-day outcomes in adult patients with acute appendicitis. Acknowledging the limitations described herein, these data might help guide the use of the potentially limited operative and professional resources allocated for emergency surgical care. Such information will contribute to the efficient allocation of limited surgical resources without negatively affecting patient care.
Correspondence: Angela M. Ingraham, MD, MS, Division of Research and Optimal Patient Care, American College of Surgeons, 633 N St Clair St, Floor 22NE, Chicago, IL 60611 (email@example.com).
Accepted for Publication: May 17, 2010.
Author Contributions: All authors 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. Study concept and design: Ingraham and Nathens. Acquisition of data: Ingraham, Ko, and Russell. Analysis and interpretation of data: Ingraham, Cohen, Bilimoria, Ko, Hall, Russell, and Nathens. Drafting of the manuscript: Ingraham and Nathens. Critical revision of the manuscript for important intellectual content: Ingraham, Cohen, Bilimoria, Ko, Hall, Russell, and Nathens. Statistical analysis: Ingraham, Cohen, Bilimoria, Hall, and Nathens. Obtained funding: Ko and Russell. Administrative, technical, and material support: Ko and Russell. Study supervision: Ko, Hall, and Nathens.
Financial Disclosure: None reported.
Funding/Support: Dr Ingraham is supported by the Clinical Scholar in Residence Program at the American College of Surgeons. Dr Bilimoria is supported by a Priority Grant from Northwestern University. Dr Nathens is supported by a Canada Research Chair in Systems of Trauma Care. Dr Hall is supported by the Center for Health Policy, Washington University in St Louis.
Previous Presentations: This paper was presented at the 81st Annual Meeting of the Pacific Coast Surgical Association; February 14, 2010; Kapalua, Hawaii; and is published after peer review and revision.