- Email: [email protected]
Outcomes of hospitalized patients undergoing emergency general surgery remote from admission
Presented at the Academic Surgical Congress 2017 Outcomes of hospitalized patients undergoing emergency general surgery remote from admission Catherine E. Sharoky, MD,a Elizabeth A. Bailey, MD, MEd,a Morgan M. Sellers, MD,a Elinore J. Kaufman, MD, MSHP,b Andrew J. Sinnamon, MD,a Christopher J. Wirtalla, BA,a Daniel N. Holena, MD, MSCE,a,c and Rachel R. Kelz, MD, MSCE,a Philadelphia, PA, and New York, NY
Background. Emergency general surgery during hospitalization has not been well characterized. We examined emergency operations remote from admission to identify predictors of postoperative 30-day mortality, postoperative duration of stay >30 days, and complications. Methods. Patients >18 years in The American College of Surgeons National Surgical Quality Improvement Program (2011–2014) who had 1 of 7 emergency operations between hospital day 3–18 were included. Patients with operations >95th percentile after admission (>18 days; n = 581) were excluded. Exploratory laparotomy only (with no secondary procedure) represented either nontherapeutic or decompressive laparotomy. Multivariable logistic regression was used to identify predictors of study outcomes. Results. Of 10,093 patients with emergency operations, most were elderly (median 66 years old [interquartile ratio: 53–77 years]), white, and female. Postoperative 30-day mortality was 12.6% (n = 1,275). Almost half the cohort (40.1%) had a complication. A small subset (6.8%) had postoperative duration of stay >30 days. Postoperative mortality after exploratory laparotomy only was particularly high (>40%). In multivariable analysis, an operation on hospital day 11–18 compared with day 3–6 was associated with death (odds ratio 1.6 [1.3–2.0]), postoperative duration of stay >30 days (odds ratio 2.0 [1.6–2.6]), and complications (odds ratio 1.5 [1.3–1.8]). Exploratory laparotomy only also was associated with death (odds ratio 5.4 [2.8–10.4]). Conclusion. Emergency general surgery performed during a hospitalization is associated with high morbidity and mortality. A longer hospital course before an emergency operation is a predictor of poor outcomes, as is undergoing exploratory laparotomy only. (Surgery 2017;162:612-9.) From the Department of Surgery,a Center for Surgery and Health Economics, Hospital of the University of Pennsylvania, Philadelphia, PA; Department of Surgery,b Weill-Cornell School of Medicine, New York, NY; and Division of Traumatology,c Surgical Critical Care and Emergency Surgery, Hospital of the University of Pennsylvania, Philadelphia, PA Supported by the University of Pennsylvania Health System and the Perelman School of Medicine at the University of Pennsylvania. Presented at the 12th Annual Academic Surgical Congress in Las Vegas, NV, February 7–9, 2017. Accepted for publication May 3, 2017. Reprint requests: Catherine E. Sharoky, MD, Department of Surgery, Hospital of the University of Pennsylvania, 3400 Spruce Street, 4 Maloney, Philadelphia, PA 19104. E-mail: catherine. [email protected]. 0039-6060/$ - see front matter Ó 2017 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.surg.2017.05.008
612 SURGERY
EMERGENCY GENERAL SURGERY (EGS) includes the management of patients with urgent and emergent nontraumatic operative conditions.1 Some EGS patients will present with an acute condition and undergo immediate operative intervention, while some may be trialed nonoperatively.2 When nonoperative management fails, these patients require operative interventions in timeframes beyond those that are typically considered in EGS studies.3,4 Other EGS patients may be hospitalized for medical disorders but develop an acute operative condition after admission which necessitates intervention. The outcomes for patients with
Surgery Volume 162, Number 3
EGS conditions are variable with high rates of postoperative adverse events.5 The complexity of the EGS population also is reflected in its strain on the health care system. EGS diseases accounted for 2.6 million hospitalizations and $28.4 billion in national costs in 2010, with costs projected to at least double by 2060.6 In the past several years much work has been done to define the scope of EGS practice and to better characterize the EGS patient cohort.1 However, most research has focused on patients who require emergency operative intervention on initial presentation (usually within 1–2 days of hospitalization),7 less is known about patients who require operative interventions remote from admission. This group, which comprises >10% of the total EGS cohort,7 may be at increased risk of poor outcomes because of the compounding effect of a hospital stay on their underlying comorbidities.8 A more complete characterization of the patients that require EGS remote from admission is important in understanding the true opportunity to improve outcomes for all EGS patients. We sought to characterize patients who underwent EGS >2 days after hospital admission. Additionally, we sought to identify factors associated with an increased risk of 30-day postoperative mortality, duration of postoperative stay >30 days, and postoperative complications in this cohort. METHODS Study population. The American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) database (2011–2014) was used to identify patients $18 years old who underwent 1 of 7 emergency general surgery operations. ACS NSQIP is a valid and reliable9 database used to assess risk factors and risk-adjusted outcomes of operative interventions from participating hospitals across the country using data extracted from medical charts by trained clinical reviewers.10 Current Procedural Terminology (CPT) code was used to identify patients who had 1 of 7 EGS procedures: appendectomy, partial colectomy, small bowel resection, operative management of peptic ulcer disease, cholecystectomy, lysis of adhesions, and laparotomy. These procedures were chosen as they were recently shown by Scott et al to represent 80% of the EGS national operative burden (deaths, complications, and costs).7 In our study, patients with a secondary EGS CPT code listed along with a primary laparotomy code were reclassified as having had the
Sharoky et al 613
second-listed procedure. Exploratory laparotomy only (ELO) then was defined as patients with a primary CPT of exploratory laparotomy with no accompanying second procedure. The ELO cohort represented patients who had either a “negative” or nontherapeutic laparotomy, or had ELO for conditions such as abdominal compartment syndrome for which laparotomy is therapeutic. Patients were included in the study if they underwent the operative procedure of interest after hospital day 2. We excluded patients >95th percentile of days from hospitalization to operative intervention (>18 days; n = 581) given the rarity of this event. Only operative cases classified as emergent were included in our analysis. An emergency case as defined by ACS NSQIP occurs shortly after the onset of symptomatology with the implication that clinical deterioration could occur unpredictably or rapidly with delay.10 Emergency status is reported by the surgeon or anesthesiologist at the time of operation. Study variables. We collected data on patient demographics, preoperative clinical variables, and postoperative clinical outcomes. Patient demographics included sex, race (white, black, other/unknown), age (<50, 50–64, 65–79, $80 years), functional status on admission, American Society of Anesthesiologists (ASA) class, and transfer status. Preoperative clinical variables included diagnosis of sepsis (ranging from systemic inflammatory response syndrome to septic shock), mechanical ventilation 48 hours prior to operation, disseminated cancer, open wound at the time of operation, and renal failure 24 hours prior to operation. Patients were initially grouped into four 4-day intervals to examine the association between study outcomes and preoperative duration of stay. Interval 3 and 4 were combined due to small sample size. The final operative intervals were day 3– 6 (n = 7,405), day 7–10 (n = 1,730), and day 11–18 (n = 958). Postoperative clinical outcomes of interest included 30-day postoperative mortality, length of postoperative stay >30 days (LOPS30), and any postoperative complication (a binary outcome signifying the presence of at least one of the following complications: operative site or organ space infection, wound dehiscence, pneumonia, unplanned reintubation, pulmonary embolism/ deep vein thrombosis, failure of extubation within 48 hours, sepsis, myocardial infarction, acute renal insufficiency, urinary tract infection, stroke, cardiac arrest, bleed requiring transfusion, unplanned reoperation). Statistical analysis. Descriptive statistics were used to characterize patients who underwent EGS
614 Sharoky et al
Surgery September 2017
Table I. Demographic and preoperative clinical characteristics of EGS patients Characteristic Demographic Age, median (IQR), y Sex female, no. (%) Race, no. (%) White Black Other/Unknown Admitted from home, no. (%) Preoperative clinical Dependent functional status, no. (%) Mechanical ventilation, no. (%) ASA class, no. (%) I and II III IV and V SIRS-septic shock, no. (%) Disseminated cancer, no. (%) Open wound, no. (%) Acute renal failure, no. (%)
EGS cohort N = 10,093 66 (53–77) 5,519 (54.7) 6,666 1,309 2,118 8,280
(66.1) (13.0) (21.0) (82.0)
1,209 (12.0) 831 (8.2) 2,507 4,386 3,186 3,973 685 595 394
(24.9) (43.5) (31.6) (39.4) (6.8) (5.9) (3.9)
SIRS, Systemic inflammatory response syndrome.
during hospitalization. Median values with interquartile ranges were reported for continuous variables that were not normally distributed, percentages were reported for categorical variables. Univariate analysis was preformed using v2 tests to identify factors associated with each of our 3 outcomes. Factors significantly associated (P < .10) with each outcome were included in multivariable analysis for that specific outcome. Independent multivariable logistic regression models were built to examine risk factors associated with study outcomes. Bonferroni correction was used to address multiple comparisons. Patients without a documented ASA class (n = 14) and without a discharge date (n = 50) were excluded from multivariable analysis. Patients who died (n = 1,275) were excluded from LOPS30 analysis. Analyses were performed using Stata/MP 13.1 statistical software (StataCorp, College Station, TX).11 This study was deemed exempt from review by the University of Pennsylvania Institutional Review Board. RESULTS Population characteristics. Our study consisted of 10,093 patients who had 1 of 7 emergency procedures between hospital day 3–18. Demographic and preoperative characteristics of the cohort are shown in Table I. More than half of
Table II. Postoperative clinical outcomes of EGS patients Postoperative outcome 30-day postoperative mortality, no. (%) Time from operation to death, median (IQR), d LOPS30, no. (%) LOPS30 duration of postoperative stay, median (IQR), d Full cohort duration of postoperative stay, median (IQR), d Postoperative complications (any), no. (%) Bleed requiring transfusion, no. (%) Wound complication,* no. (%) Failure to extubate in 48 hours, no. (%) Septic shock or sepsis, no. (%) Pneumonia, no. (%) Unplanned reintubation, no. (%) Pulmonary embolism/deep vein thrombosis, no. (%) Acute kidney injury, no. (%) Cardiac complication,y no (%)
EGS cohort N = 10,093 1,275 (12.6) 8 (2–16) 594 (6.8) 41 (35–52)
8 (IQR: 4–14)
4,096 (40.1%) 1,764 (16.5) 1,277 (12.7) 1,031 (10.2) 936 (8.8) 886 (8.8) 660 (6.2) 381 (3.8) 342 (3.4) 335 (3.2)
*Superficial/deep operative site infection, organ/deep space infection, wound dehiscence. yCardiac arrest, myocardial infarction.
patients were white (66.1%) and female (54.7%), and the median age was 66 years (interquartile ratio [IQR] 53–77 years). Most of the patients were admitted from home (82%) and were functionally independent on admission (88%). A majority were ASA Class $III (74.7%) signifying at least severe systemic disease with functional limitations at the time of operation. A significant proportion of patients (39.6%) were diagnosed with sepsis preoperatively. Outcomes. The percentage of patients who experienced each study outcome, including specific complications, are shown in Table II. The overall 30-day postoperative mortality in this cohort was 12.6% (n = 1,275). For patients who died, the median time from operation to death was 8 days (IQR 2–16 days). In multivariable analysis controlling for potential confounders, patients who had EGS on hospital day 11–18 were 60% more likely to die than patients who had EGS on hospital day 3–6 (OR 1.6 [1.3–2.0]). ASA class $III, increasing age, dependent
Sharoky et al 615
Surgery Volume 162, Number 3
Table III. Multivariable logistic regression: Predictors of 30-day postoperative mortality, LOPS30, postoperative complication in EGS patients N Hospitalization tooperation (d) 3–6 7–10 11–18 Operation Appendectomy Partial colectomy Small bowel resection Operative management of PUDz Cholecystectomy Lysis of adhesions ELO Age (y) <50 50–64 65–79 $80 Race White Black Other/Unknown ASA class I and II III IV and V Dependent functional status Not admitted from home Preoperative SIRS-septic shock Preoperative mechanical ventilation
30-day postoperative mortality
LOPS30
Postoperative complication
1,275
594
4,096
(ref) 1.2* [1.0–1.4] 1.6* [1.3–2.0]
1.9* [1.5–2.4] 2.0* [1.6–2.6]
1.4* [1.2–1.6] 1.5* [1.3–1.8]
4.4* 4.3* 3.0* 1.0 1.8 4.0*
2.6* 2.8* 2.3* 0.5* 1.0 1.4
2.1* 2.6* 1.4 0.6 1.5 5.4*
(ref) [1.1–3.8] [1.4–4.8] [0.7–3.0] [0.3–1.3] [0.8–2.8] [2.8–10.4]
[1.9–10.0] [1.8–10.0] [1.1–8.0] [0.4–2.4] [0.7–4.4] [1.6–10.0]
[2.0–3.4] [2.1–3.6] [1.5–3.4] [0.4–0.6] [0.8–1.5] [1.0–1.9]
(ref) 2.0* [1.5–2.8] 3.3* [2.4–4.4] 4.8* [3.5–6.6]
1.0 [0.8–1.4] 1.0 [0.7–1.3] 1.0 [0.8–1.5]
1.1 [1.0–1.3] 1.2* [1.1–1.4] 1.2* [1.0–1.4]
(ref) 0.7* [0.6–0.9] 0.8 [0.7–1.0]
1.4* [1.1–1.8] 2.3* [1.9–2.9]
1.1 [1.0–1.3] 1.1 [1.0–1.2]
4.2* 16.8* 1.6* 1.1 2.2* 2.3*
(ref) [2.5–7.0] [10.5, 27.9] [1.4–1.9] [0.9–1.3] [1.9–2.6] [1.9–2.8]
2.8* 4.9* 1.3* 1.4* 2.6* 3.5*
[1.9–4.2] [3.2–7.4] [1.1–1.7] [1.1–1.7] [2.1–3.2] [2.7–4.6]
2.1* 4.1* 1.4* 1.3* 2.2* 1.0
[1.9–2.5] [3.5–4.8] [1.2–1.6] [1.1–1.4] [2.0–2.4] [0.8–1.1]
Ratio: 95% confidence interval in brackets. *P < .05; bold values remained significant after Bonferroni correction. PUD, Peptic ulcer disease; SIRS, systemic inflammatory response syndrome.
functional status, sepsis, mechanical ventilation and small bowel resection also were associated with death in the multivariable model. Black race was associated with decreased odds of death (OR 0.7 [0.6–0.9]). Multivariable logistic regression models are shown in Table III. Median postoperative LOS for all patients who survived to hospital discharge was 8 days (IQR 4–14 days). A small percentage of patients (6.8%) had LOPS30. In multivariable analysis, EGS on hospital day 11–18 resulted in twice the risk of LOPS30 (OR 2.0 [1.6–2.6]) compared with EGS on day 3–6. Odds of LOPS30 was 4 times greater for patients who had a partial colectomy (OR 4.4 [1.9– 10.0]), small bowel resection (OR 4.3 [1.8–10.0]), or ELO (OR 4.0 [1.6–10.0]) compared with patients undergoing appendectomy (Table III). ASA class $III, sepsis, mechanical ventilation, and nonwhite race also were associated with LOPS30.
Almost half of the cohort experienced a postoperative complication (40.1%). The most common complications were bleed requiring transfusion (16.5%), wound complications (12.7%), and failure to extubate within 48 hours in patients without preoperative respiratory failure (10.2%) (Table II). The percentage of patients who developed a postoperative complication increased with longer time from hospital admission to operation (day 3–6: 35%; day 7–10: 53%; day 11–18: 61% P < .001) (Fig 1). In multivariable analysis, EGS on hospital day 11–18 was associated with a 50% increased odds of complications (OR 1.5 [1.3–1.8]) compared with EGS on day 3–6 (Table III). Operative predictors of complications included small bowel resection (OR 2.8 [2.1–3.6]), partial colectomy (OR 2.6 [2.0–3.4]), and operative management of peptic ulcer disease (OR 2.3 [1.5–3.4]). ASA Class $III,
616 Sharoky et al
Surgery September 2017
Fig 1. Postoperative complication rate by operative day after admission in EGS patients.
Fig 2. EGS procedure frequency and associated 30-day postoperative mortality rate.
preoperative sepsis, age, dependent functional status, and transfer status also were predictive of complications. Cholecystectomy was associated with decreased risk of complications in the model (OR 0.5 [0.4–0.7]). (Table III). Among the 7 procedures, ELO had the greatest 30-day postoperative mortality (41%; n = 376) (Fig 2). An additional 3.2% (n = 12) ELO patients died beyond the 30-day postoperative window. In multivariable analysis, patients who underwent ELO were 5 times more likely to die than patients who underwent appendectomy (OR 5.4 [2.8–10.4]). The most frequent postoperative
diagnoses for ELO patients are shown in Table IV. More than half (60%) of the ELO patients who survived to discharge (n = 208) were discharged home. The remainder (35.1%, n = 73) were discharged to a skilled facility for further care. Discharge destinations for ELO patients are shown in Table V. DISCUSSION In this study, we found that undergoing an emergency general surgery operation during an inpatient stay was associated with high rates of postoperative mortality (12.6%) and complications
Sharoky et al 617
Surgery Volume 162, Number 3
Table IV. Most frequent postoperative diagnoses for ELO patients Postoperative diagnosis Intestinal obstruction Vascular insufficiency of intestine Septicemia Nontraumatic abdominal compartment syndrome Paralytic ileus
N = 376 (%) 44 42 12 11
(11.9) (11.3) (3.2) (3.0)
9 (2.4)
Table V. Discharge destination of ELO patients who survived until hospital discharge Discharge destination Home Skilled care facility* Rehabilitation facility Transferred to another acute care facility Unknown/Missing Returned to chronic care facility Unskilled nursing facility/assisted-living
N = 208 (%) 124 45 18 10 9 1 1
(59.6) (21.6) (8.7) (4.8) (4.3) (0.5) (0.5)
*Transitional care unit, subacute hospital, ventilator bed, or skilled nursing home.
(40.1%). The risk of death and complications increased with a longer hospital course prior to operation. Patients with a longer preoperative course also had a greater risk of a postoperative stay >30 days. Outcome rates in our study population are similar to those reported in prior studies examining EGS populations,5,12,13 including those focused on EGS within 48 hours of admission.7 To our knowledge, our study is the first to show that the high morbidity and mortality associated with EGS exists regardless of the operative timeframe. Our study likely captures several types of EGS patients. One subset includes patients that have an operative diagnosis on admission and fail nonoperative management. Another includes patients being treated for a medical condition who develop an acute operative indication at some later point. A third includes patients who have an acute operative condition on presentation that is initially misdiagnosed, with poor outcomes reflecting the consequences of a delay in recognition. These patient subsets are all systematically different than EGS patients who undergo an operation on initial presentation. Patients trialed nonoperatively are often older, with more comorbidities that preclude operation than the immediate operation group.14 This is reflected in our finding that most patients in our EGS cohort were older adults (>65 years). Patients who develop an acute operative problem while hospitalized may already be in a tenuous health state due to their medical illness. This is reflected in the high percentage of patients assigned ASA Class $3 in our cohort. Hospitalization itself has been shown to deplete physiologic reserve and increase patient vulnerability,8,15 which compounded with the effects of an acute operative process and underlying comorbidities, could explain the poor postoperative outcomes we observe. We found that patients who had an operation later in their hospital course had greater odds of
all adverse outcomes. Both nutritional deficiencies and operative condition-specific factors may contribute to these findings. Malnutrition is a well-known risk factor for poor operative outcomes,16 and >60% of older hospitalized patients have protein malnutrition on admission or develop it while acutely ill due to increased metabolic needs.17 Depletion of nutritional reserve during a prolonged preoperative hospital course (particularly for patients not allowed to take anything by mouth during a nonoperative trial) could add to the negative physiologic effects seen with hospitalization.8 The natural history of specific EGS conditions also is important to consider. In appendicitis for example, studies show that rupture rates and ensuing complications increase when operation is delayed beyond 36 hours.18 In small bowel obstruction, some studies show operative delay increases the risk of resection at operation.19 While the progression of EGS conditions can be unpredictable, our finding of poor outcomes with a longer preoperative timeframe adds support to the argument that shorter nonoperative trials may be warranted for patients with EGS conditions. A key finding in our study was that almost half of the patients who underwent ELO died postoperatively. The ELO patients are clinically complex, and often require extensive resources20 both at the systems and provider level. Costs associated with EGS care are escalating, and one of the most important factors identified as contributing to high-cost hospitalization is undergoing a major operation.21 While for many EGS patients an operation is the best or only therapeutic option, the benefits in the subset represented by ELO are not always as clear. While the mortality rate after ELO in our study approaches almost 50%, for those that do survive there is a high likelihood of being discharged home. As the field moves towards goal-directed care and minimizing nonbeneficial operation,22 the dichotomy of this outcome
618 Sharoky et al
highlights the importance of discussions focused on understanding patient preferences and setting realistic postoperative expectations in clinically challenging situations like those represented by the ELO cohort. More focused research also is needed to examine factors potentially predictive of survival and home discharge in ELO patients. Limitations of our study include its retrospective nature and the inherent limitations associated with an observational study. The emergent designation of a case within ACS NSQIP is subjective, and therefore variable among providers at ACS NSQIP-participating hospitals. Previous studies testing inter-rater reliability, however, have suggested ACS-NSQIP is of sufficient quality,23 and the emergency variable has been validated by use in prior studies24 to similarly identify EGS populations. The emergency variable in ACS NSQIP allowed for the identification of emergent operations at any point during hospitalization regardless of admission classification. However, using ASC NSQIP limited our ability to characterize the different types of patients who ultimately required EGS remote from admission. Future research aimed at preoperative characterization of these patients will provide valuable insight into the clinical pathways that lead to remote EGS and may identify targets for improvements in care delivery. ACS NSQIP does not capture patient insurance or benefit status, and we were unable to detect its influence on the association between EGS and LOPS30 in this study. In addition, in limiting our study to 7 specific EGS operations we are likely underrepresenting the breadth of EGS operations performed on hospitalized patients. However, the use of a defined set of high-burden EGS operations helps to standardize and generalize EGS studies and provides a framework for further investigation.1,25 Using a national clinical dataset, we examined patients who required emergency general surgery operations remote from admission and found that the risk of postoperative death, complications and a long postoperative stay was significantly increased with a longer preoperative hospital course. Overall morbidity and mortality in the cohort was high. Almost half of the patients who underwent nontherapeutic or decompressive laparotomy died postoperatively. Our findings emphasize the importance of setting realistic postoperative expectations for patients who develop the need for emergent operative intervention remote from admission.
Surgery September 2017
REFERENCES 1. Shafi S, Aboutanos MB, Agarwal S Jr, Brown CV, Crandall M, Feliciano DV, et al. Emergency general surgery: definition and estimated burden of disease. J Trauma Acute Care Surg 2013;74:1092-7. 2. Di Saverio S, Coccolini F, Galati M, Smerieri N, Biffl WL, Ansaloni L, et al. Bologna guidelines for diagnosis and management of adhesive small bowel obstruction (ASBO): 2013 update of the evidence-based guidelines from the world society of emergency surgery ASBO working group. World J Emerg Surg 2013;8:42. 3. Schraufnagel D, Rajaee S, Millham FH. How many sunsets? Timing of surgery in adhesive small bowel obstruction: a study of the Nationwide Inpatient Sample. J Trauma Acute Care Surg 2013;74:181-7; discussion 7-9. 4. Diaz JJ Jr, Bokhari F, Mowery NT, Acosta JA, Block EF, Bromberg WJ, et al. Guidelines for management of small bowel obstruction. J Trauma 2008;64:1651-64. 5. Stonelake S, Thomson P, Suggett N. Identification of the high risk emergency surgical patient: which risk prediction model should be used? Ann Med Surg (Lond) 2015;4:240-7. 6. Ogola GO, Gale SC, Haider A, Shafi S. The financial burden of emergency general surgery: national estimates 2010 to 2060. J Trauma Acute Care Surg 2015;79:444-8. 7. Scott JW, Olufajo OA, Brat GA, Rose JA, Zogg CK, Haider AH, et al. Use of national burden to define operative emergency general surgery. JAMA Surg 2016;151:e160480. 8. Detsky AS, Krumholz HM. Reducing the trauma of hospitalization. JAMA 2014;311:2169-70. 9. Khuri SF. The NSQIP: a new frontier in surgery. Surgery 2005;138:837-43. 10. American College of Surgeons National Surgical Quality Improvement Program user guide for the ACS NSQIP participant use file (PUF). American College of Surgeons. 2011–2014. 11. StataCorp. 2013. Stata statistical software: release 13. College Station TSL. 12. Havens JM, Peetz AB, Do WS, Cooper Z, Kelly E, Askari R, et al. The excess morbidity and mortality of emergency general surgery. J Trauma Acute Care Surg 2015;78:306-11. 13. Lees MC, Merani S, Tauh K, Khadaroo RG. Perioperative factors predicting poor outcome in elderly patients following emergency general surgery: a multivariate regression analysis. Can J Surg 2015;58:312-7. 14. McGillicuddy EA, Schuster KM, Barre K, Suarez L, Hall MR, Kaml GJ, et al. Non-operative management of acute cholecystitis in the elderly. Br J Surg 2012;99:1254-61. 15. Krumholz HM. Post-hospital syndrome---an acquired, transient condition of generalized risk. N Engl J Med 2013;368:100-2. 16. Gibbs J, Cull W, Henderson W, Daley J, Hur K, Khuri SF. Preoperative serum albumin level as a predictor of operative mortality and morbidity: results from the National VA Surgical Risk Study. Arch Surg 1999;134:36-42. 17. Gary R, Fleury J. Nutritional status: key to preventing functional decline in hospitalized older adults. Topics in Geriatric Rehabilitation 2002;17:40-71. 18. Bickell NA, Aufses AH Jr, Rojas M, Bodian C. How time affects the risk of rupture in appendicitis. J Am Coll Surg 2006;202:401-6. 19. Bickell NA, Federman AD, Aufses AH Jr. Influence of time on risk of bowel resection in complete small bowel obstruction. J Am Coll Surg 2005;201:847-54.
Surgery Volume 162, Number 3
20. Ko A, Murry JS, Hoang DM, Harada MY, Aquino L, Coffey C, et al. High-value care in the surgical intensive care unit: effect on ancillary resources. J Surg Res 2016; 202:455-60. 21. Ogola GO, Shafi S. Cost of specific emergency general surgery diseases and factors associated with high-cost patients. J Trauma Acute Care Surg 2016;80:265-71. 22. Cooper Z, Courtwright A, Karlage A, Gawande A, Block S. Pitfalls in communication that lead to nonbeneficial emergency surgery in elderly patients with serious illness: description of the problem and elements of a solution. Ann Surg 2014;260:949-57.
Sharoky et al 619
23. Shiloach M, Frencher SK Jr, Steeger JE, Rowell KS, Bartzokis K, Tomeh MG, et al. Toward robust information: data quality and inter-rater reliability in the American College of Surgeons National Surgical Quality Improvement Program. J Am Coll Surg 2010;210:6-16. 24. Muthuvel G, Tevis SE, Liepert AE, Agarwal SK, Kennedy GD. A composite index for predicting readmission following emergency general surgery. J Trauma Acute Care Surg 2014;76:1467-72. 25. Paul MG. The public health crisis in emergency general surgery: who will pay the price and bear the burden? JAMA Surg 2016;151:e160640.
Background. Emergency general surgery during hospitalization has not been well characterized. We examined emergency operations remote from admission to identify predictors of postoperative 30-day mortality, postoperative duration of stay >30 days, and complications. Methods. Patients >18 years in The American College of Surgeons National Surgical Quality Improvement Program (2011–2014) who had 1 of 7 emergency operations between hospital day 3–18 were included. Patients with operations >95th percentile after admission (>18 days; n = 581) were excluded. Exploratory laparotomy only (with no secondary procedure) represented either nontherapeutic or decompressive laparotomy. Multivariable logistic regression was used to identify predictors of study outcomes. Results. Of 10,093 patients with emergency operations, most were elderly (median 66 years old [interquartile ratio: 53–77 years]), white, and female. Postoperative 30-day mortality was 12.6% (n = 1,275). Almost half the cohort (40.1%) had a complication. A small subset (6.8%) had postoperative duration of stay >30 days. Postoperative mortality after exploratory laparotomy only was particularly high (>40%). In multivariable analysis, an operation on hospital day 11–18 compared with day 3–6 was associated with death (odds ratio 1.6 [1.3–2.0]), postoperative duration of stay >30 days (odds ratio 2.0 [1.6–2.6]), and complications (odds ratio 1.5 [1.3–1.8]). Exploratory laparotomy only also was associated with death (odds ratio 5.4 [2.8–10.4]). Conclusion. Emergency general surgery performed during a hospitalization is associated with high morbidity and mortality. A longer hospital course before an emergency operation is a predictor of poor outcomes, as is undergoing exploratory laparotomy only. (Surgery 2017;162:612-9.) From the Department of Surgery,a Center for Surgery and Health Economics, Hospital of the University of Pennsylvania, Philadelphia, PA; Department of Surgery,b Weill-Cornell School of Medicine, New York, NY; and Division of Traumatology,c Surgical Critical Care and Emergency Surgery, Hospital of the University of Pennsylvania, Philadelphia, PA Supported by the University of Pennsylvania Health System and the Perelman School of Medicine at the University of Pennsylvania. Presented at the 12th Annual Academic Surgical Congress in Las Vegas, NV, February 7–9, 2017. Accepted for publication May 3, 2017. Reprint requests: Catherine E. Sharoky, MD, Department of Surgery, Hospital of the University of Pennsylvania, 3400 Spruce Street, 4 Maloney, Philadelphia, PA 19104. E-mail: catherine. [email protected]. 0039-6060/$ - see front matter Ó 2017 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.surg.2017.05.008
612 SURGERY
EMERGENCY GENERAL SURGERY (EGS) includes the management of patients with urgent and emergent nontraumatic operative conditions.1 Some EGS patients will present with an acute condition and undergo immediate operative intervention, while some may be trialed nonoperatively.2 When nonoperative management fails, these patients require operative interventions in timeframes beyond those that are typically considered in EGS studies.3,4 Other EGS patients may be hospitalized for medical disorders but develop an acute operative condition after admission which necessitates intervention. The outcomes for patients with
Surgery Volume 162, Number 3
EGS conditions are variable with high rates of postoperative adverse events.5 The complexity of the EGS population also is reflected in its strain on the health care system. EGS diseases accounted for 2.6 million hospitalizations and $28.4 billion in national costs in 2010, with costs projected to at least double by 2060.6 In the past several years much work has been done to define the scope of EGS practice and to better characterize the EGS patient cohort.1 However, most research has focused on patients who require emergency operative intervention on initial presentation (usually within 1–2 days of hospitalization),7 less is known about patients who require operative interventions remote from admission. This group, which comprises >10% of the total EGS cohort,7 may be at increased risk of poor outcomes because of the compounding effect of a hospital stay on their underlying comorbidities.8 A more complete characterization of the patients that require EGS remote from admission is important in understanding the true opportunity to improve outcomes for all EGS patients. We sought to characterize patients who underwent EGS >2 days after hospital admission. Additionally, we sought to identify factors associated with an increased risk of 30-day postoperative mortality, duration of postoperative stay >30 days, and postoperative complications in this cohort. METHODS Study population. The American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) database (2011–2014) was used to identify patients $18 years old who underwent 1 of 7 emergency general surgery operations. ACS NSQIP is a valid and reliable9 database used to assess risk factors and risk-adjusted outcomes of operative interventions from participating hospitals across the country using data extracted from medical charts by trained clinical reviewers.10 Current Procedural Terminology (CPT) code was used to identify patients who had 1 of 7 EGS procedures: appendectomy, partial colectomy, small bowel resection, operative management of peptic ulcer disease, cholecystectomy, lysis of adhesions, and laparotomy. These procedures were chosen as they were recently shown by Scott et al to represent 80% of the EGS national operative burden (deaths, complications, and costs).7 In our study, patients with a secondary EGS CPT code listed along with a primary laparotomy code were reclassified as having had the
Sharoky et al 613
second-listed procedure. Exploratory laparotomy only (ELO) then was defined as patients with a primary CPT of exploratory laparotomy with no accompanying second procedure. The ELO cohort represented patients who had either a “negative” or nontherapeutic laparotomy, or had ELO for conditions such as abdominal compartment syndrome for which laparotomy is therapeutic. Patients were included in the study if they underwent the operative procedure of interest after hospital day 2. We excluded patients >95th percentile of days from hospitalization to operative intervention (>18 days; n = 581) given the rarity of this event. Only operative cases classified as emergent were included in our analysis. An emergency case as defined by ACS NSQIP occurs shortly after the onset of symptomatology with the implication that clinical deterioration could occur unpredictably or rapidly with delay.10 Emergency status is reported by the surgeon or anesthesiologist at the time of operation. Study variables. We collected data on patient demographics, preoperative clinical variables, and postoperative clinical outcomes. Patient demographics included sex, race (white, black, other/unknown), age (<50, 50–64, 65–79, $80 years), functional status on admission, American Society of Anesthesiologists (ASA) class, and transfer status. Preoperative clinical variables included diagnosis of sepsis (ranging from systemic inflammatory response syndrome to septic shock), mechanical ventilation 48 hours prior to operation, disseminated cancer, open wound at the time of operation, and renal failure 24 hours prior to operation. Patients were initially grouped into four 4-day intervals to examine the association between study outcomes and preoperative duration of stay. Interval 3 and 4 were combined due to small sample size. The final operative intervals were day 3– 6 (n = 7,405), day 7–10 (n = 1,730), and day 11–18 (n = 958). Postoperative clinical outcomes of interest included 30-day postoperative mortality, length of postoperative stay >30 days (LOPS30), and any postoperative complication (a binary outcome signifying the presence of at least one of the following complications: operative site or organ space infection, wound dehiscence, pneumonia, unplanned reintubation, pulmonary embolism/ deep vein thrombosis, failure of extubation within 48 hours, sepsis, myocardial infarction, acute renal insufficiency, urinary tract infection, stroke, cardiac arrest, bleed requiring transfusion, unplanned reoperation). Statistical analysis. Descriptive statistics were used to characterize patients who underwent EGS
614 Sharoky et al
Surgery September 2017
Table I. Demographic and preoperative clinical characteristics of EGS patients Characteristic Demographic Age, median (IQR), y Sex female, no. (%) Race, no. (%) White Black Other/Unknown Admitted from home, no. (%) Preoperative clinical Dependent functional status, no. (%) Mechanical ventilation, no. (%) ASA class, no. (%) I and II III IV and V SIRS-septic shock, no. (%) Disseminated cancer, no. (%) Open wound, no. (%) Acute renal failure, no. (%)
EGS cohort N = 10,093 66 (53–77) 5,519 (54.7) 6,666 1,309 2,118 8,280
(66.1) (13.0) (21.0) (82.0)
1,209 (12.0) 831 (8.2) 2,507 4,386 3,186 3,973 685 595 394
(24.9) (43.5) (31.6) (39.4) (6.8) (5.9) (3.9)
SIRS, Systemic inflammatory response syndrome.
during hospitalization. Median values with interquartile ranges were reported for continuous variables that were not normally distributed, percentages were reported for categorical variables. Univariate analysis was preformed using v2 tests to identify factors associated with each of our 3 outcomes. Factors significantly associated (P < .10) with each outcome were included in multivariable analysis for that specific outcome. Independent multivariable logistic regression models were built to examine risk factors associated with study outcomes. Bonferroni correction was used to address multiple comparisons. Patients without a documented ASA class (n = 14) and without a discharge date (n = 50) were excluded from multivariable analysis. Patients who died (n = 1,275) were excluded from LOPS30 analysis. Analyses were performed using Stata/MP 13.1 statistical software (StataCorp, College Station, TX).11 This study was deemed exempt from review by the University of Pennsylvania Institutional Review Board. RESULTS Population characteristics. Our study consisted of 10,093 patients who had 1 of 7 emergency procedures between hospital day 3–18. Demographic and preoperative characteristics of the cohort are shown in Table I. More than half of
Table II. Postoperative clinical outcomes of EGS patients Postoperative outcome 30-day postoperative mortality, no. (%) Time from operation to death, median (IQR), d LOPS30, no. (%) LOPS30 duration of postoperative stay, median (IQR), d Full cohort duration of postoperative stay, median (IQR), d Postoperative complications (any), no. (%) Bleed requiring transfusion, no. (%) Wound complication,* no. (%) Failure to extubate in 48 hours, no. (%) Septic shock or sepsis, no. (%) Pneumonia, no. (%) Unplanned reintubation, no. (%) Pulmonary embolism/deep vein thrombosis, no. (%) Acute kidney injury, no. (%) Cardiac complication,y no (%)
EGS cohort N = 10,093 1,275 (12.6) 8 (2–16) 594 (6.8) 41 (35–52)
8 (IQR: 4–14)
4,096 (40.1%) 1,764 (16.5) 1,277 (12.7) 1,031 (10.2) 936 (8.8) 886 (8.8) 660 (6.2) 381 (3.8) 342 (3.4) 335 (3.2)
*Superficial/deep operative site infection, organ/deep space infection, wound dehiscence. yCardiac arrest, myocardial infarction.
patients were white (66.1%) and female (54.7%), and the median age was 66 years (interquartile ratio [IQR] 53–77 years). Most of the patients were admitted from home (82%) and were functionally independent on admission (88%). A majority were ASA Class $III (74.7%) signifying at least severe systemic disease with functional limitations at the time of operation. A significant proportion of patients (39.6%) were diagnosed with sepsis preoperatively. Outcomes. The percentage of patients who experienced each study outcome, including specific complications, are shown in Table II. The overall 30-day postoperative mortality in this cohort was 12.6% (n = 1,275). For patients who died, the median time from operation to death was 8 days (IQR 2–16 days). In multivariable analysis controlling for potential confounders, patients who had EGS on hospital day 11–18 were 60% more likely to die than patients who had EGS on hospital day 3–6 (OR 1.6 [1.3–2.0]). ASA class $III, increasing age, dependent
Sharoky et al 615
Surgery Volume 162, Number 3
Table III. Multivariable logistic regression: Predictors of 30-day postoperative mortality, LOPS30, postoperative complication in EGS patients N Hospitalization tooperation (d) 3–6 7–10 11–18 Operation Appendectomy Partial colectomy Small bowel resection Operative management of PUDz Cholecystectomy Lysis of adhesions ELO Age (y) <50 50–64 65–79 $80 Race White Black Other/Unknown ASA class I and II III IV and V Dependent functional status Not admitted from home Preoperative SIRS-septic shock Preoperative mechanical ventilation
30-day postoperative mortality
LOPS30
Postoperative complication
1,275
594
4,096
(ref) 1.2* [1.0–1.4] 1.6* [1.3–2.0]
1.9* [1.5–2.4] 2.0* [1.6–2.6]
1.4* [1.2–1.6] 1.5* [1.3–1.8]
4.4* 4.3* 3.0* 1.0 1.8 4.0*
2.6* 2.8* 2.3* 0.5* 1.0 1.4
2.1* 2.6* 1.4 0.6 1.5 5.4*
(ref) [1.1–3.8] [1.4–4.8] [0.7–3.0] [0.3–1.3] [0.8–2.8] [2.8–10.4]
[1.9–10.0] [1.8–10.0] [1.1–8.0] [0.4–2.4] [0.7–4.4] [1.6–10.0]
[2.0–3.4] [2.1–3.6] [1.5–3.4] [0.4–0.6] [0.8–1.5] [1.0–1.9]
(ref) 2.0* [1.5–2.8] 3.3* [2.4–4.4] 4.8* [3.5–6.6]
1.0 [0.8–1.4] 1.0 [0.7–1.3] 1.0 [0.8–1.5]
1.1 [1.0–1.3] 1.2* [1.1–1.4] 1.2* [1.0–1.4]
(ref) 0.7* [0.6–0.9] 0.8 [0.7–1.0]
1.4* [1.1–1.8] 2.3* [1.9–2.9]
1.1 [1.0–1.3] 1.1 [1.0–1.2]
4.2* 16.8* 1.6* 1.1 2.2* 2.3*
(ref) [2.5–7.0] [10.5, 27.9] [1.4–1.9] [0.9–1.3] [1.9–2.6] [1.9–2.8]
2.8* 4.9* 1.3* 1.4* 2.6* 3.5*
[1.9–4.2] [3.2–7.4] [1.1–1.7] [1.1–1.7] [2.1–3.2] [2.7–4.6]
2.1* 4.1* 1.4* 1.3* 2.2* 1.0
[1.9–2.5] [3.5–4.8] [1.2–1.6] [1.1–1.4] [2.0–2.4] [0.8–1.1]
Ratio: 95% confidence interval in brackets. *P < .05; bold values remained significant after Bonferroni correction. PUD, Peptic ulcer disease; SIRS, systemic inflammatory response syndrome.
functional status, sepsis, mechanical ventilation and small bowel resection also were associated with death in the multivariable model. Black race was associated with decreased odds of death (OR 0.7 [0.6–0.9]). Multivariable logistic regression models are shown in Table III. Median postoperative LOS for all patients who survived to hospital discharge was 8 days (IQR 4–14 days). A small percentage of patients (6.8%) had LOPS30. In multivariable analysis, EGS on hospital day 11–18 resulted in twice the risk of LOPS30 (OR 2.0 [1.6–2.6]) compared with EGS on day 3–6. Odds of LOPS30 was 4 times greater for patients who had a partial colectomy (OR 4.4 [1.9– 10.0]), small bowel resection (OR 4.3 [1.8–10.0]), or ELO (OR 4.0 [1.6–10.0]) compared with patients undergoing appendectomy (Table III). ASA class $III, sepsis, mechanical ventilation, and nonwhite race also were associated with LOPS30.
Almost half of the cohort experienced a postoperative complication (40.1%). The most common complications were bleed requiring transfusion (16.5%), wound complications (12.7%), and failure to extubate within 48 hours in patients without preoperative respiratory failure (10.2%) (Table II). The percentage of patients who developed a postoperative complication increased with longer time from hospital admission to operation (day 3–6: 35%; day 7–10: 53%; day 11–18: 61% P < .001) (Fig 1). In multivariable analysis, EGS on hospital day 11–18 was associated with a 50% increased odds of complications (OR 1.5 [1.3–1.8]) compared with EGS on day 3–6 (Table III). Operative predictors of complications included small bowel resection (OR 2.8 [2.1–3.6]), partial colectomy (OR 2.6 [2.0–3.4]), and operative management of peptic ulcer disease (OR 2.3 [1.5–3.4]). ASA Class $III,
616 Sharoky et al
Surgery September 2017
Fig 1. Postoperative complication rate by operative day after admission in EGS patients.
Fig 2. EGS procedure frequency and associated 30-day postoperative mortality rate.
preoperative sepsis, age, dependent functional status, and transfer status also were predictive of complications. Cholecystectomy was associated with decreased risk of complications in the model (OR 0.5 [0.4–0.7]). (Table III). Among the 7 procedures, ELO had the greatest 30-day postoperative mortality (41%; n = 376) (Fig 2). An additional 3.2% (n = 12) ELO patients died beyond the 30-day postoperative window. In multivariable analysis, patients who underwent ELO were 5 times more likely to die than patients who underwent appendectomy (OR 5.4 [2.8–10.4]). The most frequent postoperative
diagnoses for ELO patients are shown in Table IV. More than half (60%) of the ELO patients who survived to discharge (n = 208) were discharged home. The remainder (35.1%, n = 73) were discharged to a skilled facility for further care. Discharge destinations for ELO patients are shown in Table V. DISCUSSION In this study, we found that undergoing an emergency general surgery operation during an inpatient stay was associated with high rates of postoperative mortality (12.6%) and complications
Sharoky et al 617
Surgery Volume 162, Number 3
Table IV. Most frequent postoperative diagnoses for ELO patients Postoperative diagnosis Intestinal obstruction Vascular insufficiency of intestine Septicemia Nontraumatic abdominal compartment syndrome Paralytic ileus
N = 376 (%) 44 42 12 11
(11.9) (11.3) (3.2) (3.0)
9 (2.4)
Table V. Discharge destination of ELO patients who survived until hospital discharge Discharge destination Home Skilled care facility* Rehabilitation facility Transferred to another acute care facility Unknown/Missing Returned to chronic care facility Unskilled nursing facility/assisted-living
N = 208 (%) 124 45 18 10 9 1 1
(59.6) (21.6) (8.7) (4.8) (4.3) (0.5) (0.5)
*Transitional care unit, subacute hospital, ventilator bed, or skilled nursing home.
(40.1%). The risk of death and complications increased with a longer hospital course prior to operation. Patients with a longer preoperative course also had a greater risk of a postoperative stay >30 days. Outcome rates in our study population are similar to those reported in prior studies examining EGS populations,5,12,13 including those focused on EGS within 48 hours of admission.7 To our knowledge, our study is the first to show that the high morbidity and mortality associated with EGS exists regardless of the operative timeframe. Our study likely captures several types of EGS patients. One subset includes patients that have an operative diagnosis on admission and fail nonoperative management. Another includes patients being treated for a medical condition who develop an acute operative indication at some later point. A third includes patients who have an acute operative condition on presentation that is initially misdiagnosed, with poor outcomes reflecting the consequences of a delay in recognition. These patient subsets are all systematically different than EGS patients who undergo an operation on initial presentation. Patients trialed nonoperatively are often older, with more comorbidities that preclude operation than the immediate operation group.14 This is reflected in our finding that most patients in our EGS cohort were older adults (>65 years). Patients who develop an acute operative problem while hospitalized may already be in a tenuous health state due to their medical illness. This is reflected in the high percentage of patients assigned ASA Class $3 in our cohort. Hospitalization itself has been shown to deplete physiologic reserve and increase patient vulnerability,8,15 which compounded with the effects of an acute operative process and underlying comorbidities, could explain the poor postoperative outcomes we observe. We found that patients who had an operation later in their hospital course had greater odds of
all adverse outcomes. Both nutritional deficiencies and operative condition-specific factors may contribute to these findings. Malnutrition is a well-known risk factor for poor operative outcomes,16 and >60% of older hospitalized patients have protein malnutrition on admission or develop it while acutely ill due to increased metabolic needs.17 Depletion of nutritional reserve during a prolonged preoperative hospital course (particularly for patients not allowed to take anything by mouth during a nonoperative trial) could add to the negative physiologic effects seen with hospitalization.8 The natural history of specific EGS conditions also is important to consider. In appendicitis for example, studies show that rupture rates and ensuing complications increase when operation is delayed beyond 36 hours.18 In small bowel obstruction, some studies show operative delay increases the risk of resection at operation.19 While the progression of EGS conditions can be unpredictable, our finding of poor outcomes with a longer preoperative timeframe adds support to the argument that shorter nonoperative trials may be warranted for patients with EGS conditions. A key finding in our study was that almost half of the patients who underwent ELO died postoperatively. The ELO patients are clinically complex, and often require extensive resources20 both at the systems and provider level. Costs associated with EGS care are escalating, and one of the most important factors identified as contributing to high-cost hospitalization is undergoing a major operation.21 While for many EGS patients an operation is the best or only therapeutic option, the benefits in the subset represented by ELO are not always as clear. While the mortality rate after ELO in our study approaches almost 50%, for those that do survive there is a high likelihood of being discharged home. As the field moves towards goal-directed care and minimizing nonbeneficial operation,22 the dichotomy of this outcome
618 Sharoky et al
highlights the importance of discussions focused on understanding patient preferences and setting realistic postoperative expectations in clinically challenging situations like those represented by the ELO cohort. More focused research also is needed to examine factors potentially predictive of survival and home discharge in ELO patients. Limitations of our study include its retrospective nature and the inherent limitations associated with an observational study. The emergent designation of a case within ACS NSQIP is subjective, and therefore variable among providers at ACS NSQIP-participating hospitals. Previous studies testing inter-rater reliability, however, have suggested ACS-NSQIP is of sufficient quality,23 and the emergency variable has been validated by use in prior studies24 to similarly identify EGS populations. The emergency variable in ACS NSQIP allowed for the identification of emergent operations at any point during hospitalization regardless of admission classification. However, using ASC NSQIP limited our ability to characterize the different types of patients who ultimately required EGS remote from admission. Future research aimed at preoperative characterization of these patients will provide valuable insight into the clinical pathways that lead to remote EGS and may identify targets for improvements in care delivery. ACS NSQIP does not capture patient insurance or benefit status, and we were unable to detect its influence on the association between EGS and LOPS30 in this study. In addition, in limiting our study to 7 specific EGS operations we are likely underrepresenting the breadth of EGS operations performed on hospitalized patients. However, the use of a defined set of high-burden EGS operations helps to standardize and generalize EGS studies and provides a framework for further investigation.1,25 Using a national clinical dataset, we examined patients who required emergency general surgery operations remote from admission and found that the risk of postoperative death, complications and a long postoperative stay was significantly increased with a longer preoperative hospital course. Overall morbidity and mortality in the cohort was high. Almost half of the patients who underwent nontherapeutic or decompressive laparotomy died postoperatively. Our findings emphasize the importance of setting realistic postoperative expectations for patients who develop the need for emergent operative intervention remote from admission.
Surgery September 2017
REFERENCES 1. Shafi S, Aboutanos MB, Agarwal S Jr, Brown CV, Crandall M, Feliciano DV, et al. Emergency general surgery: definition and estimated burden of disease. J Trauma Acute Care Surg 2013;74:1092-7. 2. Di Saverio S, Coccolini F, Galati M, Smerieri N, Biffl WL, Ansaloni L, et al. Bologna guidelines for diagnosis and management of adhesive small bowel obstruction (ASBO): 2013 update of the evidence-based guidelines from the world society of emergency surgery ASBO working group. World J Emerg Surg 2013;8:42. 3. Schraufnagel D, Rajaee S, Millham FH. How many sunsets? Timing of surgery in adhesive small bowel obstruction: a study of the Nationwide Inpatient Sample. J Trauma Acute Care Surg 2013;74:181-7; discussion 7-9. 4. Diaz JJ Jr, Bokhari F, Mowery NT, Acosta JA, Block EF, Bromberg WJ, et al. Guidelines for management of small bowel obstruction. J Trauma 2008;64:1651-64. 5. Stonelake S, Thomson P, Suggett N. Identification of the high risk emergency surgical patient: which risk prediction model should be used? Ann Med Surg (Lond) 2015;4:240-7. 6. Ogola GO, Gale SC, Haider A, Shafi S. The financial burden of emergency general surgery: national estimates 2010 to 2060. J Trauma Acute Care Surg 2015;79:444-8. 7. Scott JW, Olufajo OA, Brat GA, Rose JA, Zogg CK, Haider AH, et al. Use of national burden to define operative emergency general surgery. JAMA Surg 2016;151:e160480. 8. Detsky AS, Krumholz HM. Reducing the trauma of hospitalization. JAMA 2014;311:2169-70. 9. Khuri SF. The NSQIP: a new frontier in surgery. Surgery 2005;138:837-43. 10. American College of Surgeons National Surgical Quality Improvement Program user guide for the ACS NSQIP participant use file (PUF). American College of Surgeons. 2011–2014. 11. StataCorp. 2013. Stata statistical software: release 13. College Station TSL. 12. Havens JM, Peetz AB, Do WS, Cooper Z, Kelly E, Askari R, et al. The excess morbidity and mortality of emergency general surgery. J Trauma Acute Care Surg 2015;78:306-11. 13. Lees MC, Merani S, Tauh K, Khadaroo RG. Perioperative factors predicting poor outcome in elderly patients following emergency general surgery: a multivariate regression analysis. Can J Surg 2015;58:312-7. 14. McGillicuddy EA, Schuster KM, Barre K, Suarez L, Hall MR, Kaml GJ, et al. Non-operative management of acute cholecystitis in the elderly. Br J Surg 2012;99:1254-61. 15. Krumholz HM. Post-hospital syndrome---an acquired, transient condition of generalized risk. N Engl J Med 2013;368:100-2. 16. Gibbs J, Cull W, Henderson W, Daley J, Hur K, Khuri SF. Preoperative serum albumin level as a predictor of operative mortality and morbidity: results from the National VA Surgical Risk Study. Arch Surg 1999;134:36-42. 17. Gary R, Fleury J. Nutritional status: key to preventing functional decline in hospitalized older adults. Topics in Geriatric Rehabilitation 2002;17:40-71. 18. Bickell NA, Aufses AH Jr, Rojas M, Bodian C. How time affects the risk of rupture in appendicitis. J Am Coll Surg 2006;202:401-6. 19. Bickell NA, Federman AD, Aufses AH Jr. Influence of time on risk of bowel resection in complete small bowel obstruction. J Am Coll Surg 2005;201:847-54.
Surgery Volume 162, Number 3
20. Ko A, Murry JS, Hoang DM, Harada MY, Aquino L, Coffey C, et al. High-value care in the surgical intensive care unit: effect on ancillary resources. J Surg Res 2016; 202:455-60. 21. Ogola GO, Shafi S. Cost of specific emergency general surgery diseases and factors associated with high-cost patients. J Trauma Acute Care Surg 2016;80:265-71. 22. Cooper Z, Courtwright A, Karlage A, Gawande A, Block S. Pitfalls in communication that lead to nonbeneficial emergency surgery in elderly patients with serious illness: description of the problem and elements of a solution. Ann Surg 2014;260:949-57.
Sharoky et al 619
23. Shiloach M, Frencher SK Jr, Steeger JE, Rowell KS, Bartzokis K, Tomeh MG, et al. Toward robust information: data quality and inter-rater reliability in the American College of Surgeons National Surgical Quality Improvement Program. J Am Coll Surg 2010;210:6-16. 24. Muthuvel G, Tevis SE, Liepert AE, Agarwal SK, Kennedy GD. A composite index for predicting readmission following emergency general surgery. J Trauma Acute Care Surg 2014;76:1467-72. 25. Paul MG. The public health crisis in emergency general surgery: who will pay the price and bear the burden? JAMA Surg 2016;151:e160640.