- Email: [email protected]
Outcomes of abdominal surgery in patients receiving mechanical ventilation for more than 48 hours
Accepted Manuscript Outcomes of abdominal surgery in patients receiving mechanical ventilation for more than 48 hours Ning Lu, M.D., Ashley Marumoto, M.D., Linda L. Wong, M.D. PII:
S0002-9610(16)30197-0
DOI:
10.1016/j.amjsurg.2015.12.029
Reference:
AJS 11891
To appear in:
The American Journal of Surgery
Received Date: 29 July 2015 Revised Date:
5 December 2015
Accepted Date: 10 December 2015
Please cite this article as: Lu N, Marumoto A, Wong LL, Outcomes of abdominal surgery in patients receiving mechanical ventilation for more than 48 hours, The American Journal of Surgery (2016), doi: 10.1016/j.amjsurg.2015.12.029. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Abstract Background Abdominal surgery in critically-ill patients has high mortality; contributing to high U.S.
RI PT
healthcare costs. This study sought to identify specific predictors of mortality in this population.
Methods Using the National Surgical Quality Improvement Program database 2006–2012, we
SC
identified 4901 patients who were intubated for more than 48 hours prior to undergoing common abdominal procedures. Mortality and predictors of mortality were determined using chi-
M AN U
square/regression analysis.
Results Overall 30-day mortality was 44.2% with increasing mortality for additional procedures performed. Ventilated patients with the following pre-operative risk factors were two to three
TE D
times as likely to die within 30 days of surgery: age greater than 65 years-old, coma, pre-op INR > 3.0, esophageal varices, and disseminated cancer.
EP
Conclusions Mortality is significant in ventilated patients who undergo abdominal surgery and is especially high with advanced age, disseminated cancer, and complications of liver disease.
AC C
Physicians should carefully discuss this with patients/family and consider palliative options when appropriate.
ACCEPTED MANUSCRIPT
Keywords Abdominal surgery Critical care
RI PT
Acute abdomen NSQIP
AC C
EP
TE D
M AN U
SC
Mortality
ACCEPTED MANUSCRIPT
Summary for the table of contents Abdominal surgery in critically-ill patients has high mortality; contributing to high U.S. healthcare costs. Using the National Surgical Quality Improvement Program database 2006–2012, we
RI PT
identified 4901 patients who were intubated for more than 48 hours prior to undergoing common abdominal procedures. Overall 30-day mortality was 44.2% with increasing mortality for additional procedures performed. Mortality is significant in ventilated patients who undergo abdominal
SC
surgery and is especially high with advanced age, disseminated cancer, and complications of liver
AC C
EP
TE D
M AN U
disease
ACCEPTED MANUSCRIPT
Outcomes of abdominal surgery in patients receiving mechanical ventilation for more than
RI PT
48 hours
Ning Lu, M.D.a
SC
Email: [email protected] Ashley Marumoto, M.D.a
Linda L. Wong, M.D.a, b
aDepartment
of Surgery, University of Hawaii, John A. Burns School of Medicine,
Honolulu, Hawaii 96813
TE D
1356 Lusitana Street, Sixth Floor
b Cancer
M AN U
Email: [email protected]
Center, University of Hawaii
701 Ilalo Street, Suite 600
EP
Honolulu, Hawai'i 96813
AC C
Corresponding author:
Linda L. Wong MD
550 South Beretania St, Suite 403 Honolulu, Hawaii 96813 Phone: (808)-523-5033 FAX: (808)-528-4940 E mail: [email protected]
1
ACCEPTED MANUSCRIPT
Introduction Critical care is expensive and has high mortality with 20% of all deaths in the United States occurring during or shortly after an intensive care unit (ICU) admission [1]. In 2000, critical care
RI PT
medicine accounted for 13.3% of hospital costs and 0.56% of the gross national domestic product in the United States [2]. Of the $585.7 billion Medicare spent in 2013, 25-30% was used towards the 5% of Medicare beneficiaries who died, which was six times the cost of Medicare spending for a
SC
survivor [3-5]. The first two days of intensive care ICU admission accounted for the highest costs, with mechanical ventilation resulting in higher daily costs and increasing the cost 2.5 fold
M AN U
compared to non-ventilated patients [6].
Admission to the ICU may be considered a therapeutic trial, whereby initial aggressive measures are transitioned to palliative once it is clear that meaningful outcomes cannot be achieved [7]. Although there is no formal definition for “futile care”, it is generally understood as care that cannot
TE D
achieve a patient’s quality of life goals. Futile care is often scrutinized for utilizing much of healthcare resources with little return. In 1995, The Study to Understand Prognosis and Preferences for Outcomes and Risk of Treatment (SUPPORT) [8,9] attempted to identify patients
EP
with poor survival prognoses and intervene so as to reduce futile care and interventions in these patients. Despite the availability of prognostic information and physician efforts to communicate
AC C
this to patients and their families, there was no reduction of futile care [8,9]. Although there is no objective means of prospectively identifying patients whose care will be futile, physicians should consider palliative care options and be sensitive to cost and utilization of resources.
Abdominal surgery is particularly risky in critically ill patients. These operations may contribute to the length of ICU admission, length of hospital stay and high cost of healthcare, but may have limited impact on survival or quality of life. A few studies have addressed mortality following
3
ACCEPTED MANUSCRIPT
exploratory laparotomy in critically ill patients but these were small samples, single center and before the era of evidence-based medicine in large databases such as the American College of Surgeons’ National Surgical Quality Improvement Program (NSQIP) [10]. Surgical risk calculators
RI PT
have been developed and are useful in estimating an individual mortality based on procedure and risk factors [11-13]. This study attempts to provide evidence-based data that can be more broadly generalized on mortality rates and predictors of mortality in critically ill patients compared to prior
SC
single-center studies. It also hopes to supplement the current NSQIP calculator by providing additional data on pre-operative risk factors and post-operative mortality that will provide
M AN U
physicians with prognostic values that can be utilized on a daily basis to help patients and their families make difficult management decisions regarding their care in the ICU.
Materials and Methods
This study is a retrospective review of the prospectively collected NSQIP database from 2006
TE D
through 2012. The database is compiled from 373 hospitals nationwide and provides information on 30-day, risk-adjusted surgical outcomes, which allows hospitals across the nation to compare outcomes. Pre-operative through 30-day post-operative variables are collected on randomly
EP
assigned patients, including patient demographics, surgical profile, preoperative risk assessment, laboratory values, operative information, and 30-day morbidity and mortality rates. A highly
AC C
trained Surgical Clinical Reviewer collects the data. All reviewers receive extensive initial training prior to starting data collection and ongoing training via continuing education. NSQIP monitors accrual rates and data sampling methodologies and conducts audits on a random basis, ensuring highly reliable data.
Patients: Common Procedural Terminology (CPT) codes [14] for frequently performed emergent abdominal procedures were used to identify patients from the NSQIP database: 49000, 43632,
4
ACCEPTED MANUSCRIPT
43635, 43621, 47600, 47562, 44120, 44125, 44950, 44960, 44970, 44143, 44320, 44140, 44150, 44204. During the time period from January 1, 2006 thru December 31, 2012, NSQIP collected data on 2,320,898 patients. Increasing numbers of patients were identified annually with 152,468
RI PT
patients recorded in 2006, the year of NSQIP’s inception and 543,885 patients added in 2012. Note that the 2006 NSQIP database did include a small number of cases that underwent a surgical
procedure in 2005 but data from the entirety of 2005 is not represented in this file. Because NSQIP
SC
does not identify whether patients are located in an ICU or on a general ward at the time of surgery, we used the parameter of “ventilated in the last 48 hours” as a surrogate for being located in an
M AN U
intensive care unit. We thus identified 17,679 (0.7%) patients who were on the ventilator for at least 48 hours and 4,901 of these patients underwent abdominal surgery as defined by the CPT codes. This was the cohort used in the study and our definition of a critically ill patient undergoing abdominal surgery. This analysis would potentially exclude patients who were critically ill but not intubated and those who were intubated for less than 48 hours prior to surgery. Trauma patients
TE D
are also excluded from the NSQIP database. For this study, we identified patients who had one of the primary CPT codes for abdominal surgery and categorized them by organ system: cholecystectomy, appendectomy, gastric, colon or small bowel procedures. We then looked at the
EP
subsequent 3 additional CPT codes to determine if other abdominal procedures were performed at this surgery. Patients were categorized as to having received exploratory laparotomy alone, one
AC C
abdominal procedure, two abdominal procedures, or three or more abdominal procedures.
NSQIP data [10] included demographics: age, sex, ethnicity, height, weight, and BMI. Comorbidities included diabetes, smoking, alcohol use, chronic obstructive pulmonary disease (COPD), concurrent pneumonia, ascites, congestive heart failure (CHF), esophageal varices, myocardial infarction within 6 months, previous percutaneous coronary intervention, angina, hypertension requiring medications, peripheral vascular disease, renal failure, dialysis, coma, cerebrovascular accident,
5
ACCEPTED MANUSCRIPT
paralysis, wound infection, disseminated cancer, steroid use, significant weight loss (10%) in the previous 6 months, bleeding disorder, prior chemotherapy, radiation, previous sepsis, systemic
RI PT
inflammatory response syndrome (SIRS), or septic shock.
Pre-operative laboratory values were divided into categorical variables as defined by: abnormal serum sodium (less than or equal to 130 or greater than or equal to 150 mmol/L) , serum
SC
creatinine greater than or equal to 2.0 mg/dL, albumin less than or equal to 3.0 gm/dL, serum bilirubin greater than or equal to 2.0 mg/dL, AST greater than 200 IU/L, alkaline phosphatase
M AN U
greater than 150 IU/L, abnormal WBC (less than 3 or greater than 12 x 109/L), hematocrit less than 30%, platelet count less than 100 x 109/L, and INR greater than 3.
Post-operative complications included superficial or deep wound infection, wound dehiscence, pneumonia, re-intubation, pulmonary embolism, failure to wean from ventilator in 48 hours, renal
TE D
insufficiency, acute renal failure, cerebrovascular accident, coma, cardiac arrest, myocardial infarction, transfusions, sepsis, septic shock and return to the operating room. Outcome variables included time in the operating room, number of units of blood transfused, return to the operating
EP
room, hospital length of stay, and 30-day mortality.
AC C
Data Endpoints: The goals of the study were to identify pre-operative risk factors for intraoperative and post-operative mortality in critically ill patients who underwent emergency abdominal surgery. NSQIP provided information on length of stay and days from operation to death. This information was used to identify patients who had died within 30-days of their procedure.
6
ACCEPTED MANUSCRIPT
Data Analysis: All analyses were performed using SPSS statistical software. Demographics (age greater than 65 years, gender), comorbidities and preoperative laboratory studies based on the categorical variables above were compared between 30-day survivors and 30-day non-survivors
RI PT
using the chi-square test. With this univariate analysis, all variables in which p <0.10 were entered into multivariate analysis to identify significant variables at p<0.05. Logistic regression was then used to calculate odds-ratios (OR) and 95% confidence intervals (CI) to identify the best predictors
SC
of 30-day mortality.
M AN U
Results
In the overall cohort of 4,901 critically ill patients who underwent emergency abdominal surgery, 53% were male, mean age was 63.9 years, and 2,560 (52.2%) were over the age of 65 years. Other characteristics are in Table 1. Preoperative laboratory studies by category included: abnormal sodium (9.6%), creatinine greater than or equal to 2.0 mg/dL (42.8%), albumin less than or equal
TE D
to 3.0 gm/dL (73.4%), bilirubin greater than or equal to 2.0 mg/dL (21.1%), AST more than 200 IU/L (9%) , alkaline phosphatase greater than 150 IU/L (16.2%), abnormal WBC (55.6%),
EP
hematocrit less than 30% (42.5%) , and platelet count less than 100 x 109/L (22.0%).
The most common procedures were those relating to the colon (1881 patients, 38.4%) followed by
AC C
exploratory laparotomy (1570 patients, 32.0%). Other primary procedures were distributed by type (Table 2). Of the 1570 patients who had exploratory laparotomy listed as the primary procedure, 1521 patients did not undergo any other abdominal procedure while 49 patients had an additional abdominal procedure performed. The 1521 patients that had only an exploratory laparotomy had a 30-day post-operative mortality rate at 51.5%. For those who were coded for 1, 2, or 3 additional abdominal procedures, the mortality rate was 39.8%, 50% and 54.2% respectively (Table 2).
7
ACCEPTED MANUSCRIPT
The most common complications included failure to wean from the ventilator within 48 hours (57.4%), transfusion requirement (26.6%), pneumonia (16.9%), septic shock (14.2%), re-
RI PT
intubation (11%), and acute renal failure (10.4%). Other complications are listed in table 3.
The major outcome measures are as follows (Table 4): In this cohort, 26.6% patients required
SC
blood transfusion and mean number of transfusions was 1.2 units (SD 2.5 units). Mean operative time was 103.7 minutes (SD 68.3 minutes) and 28% required a return to the operating room after
M AN U
the initial operation. Mean hospital length of stay was 25.8 days (SD 27.7 days). Of the 2165 patients who died, the time from operation to death was a mean of 6.75 days (range 0-30 days, SD 7.92 days). Of the survivors mean hospital length of stay following the procedure was 18.8 days.
TE D
The pre-operative factors best predicting 30-day mortality included: coma (OR 2.63, 95% CI 1.953.55), pre-operative INR greater than 3 (OR 2.6, 95% CI 1.79-3.79), esophageal varices (OR 2.44, 95% CI 1.21-4.91), and disseminated cancer (OR 1.98, 95% CI 1.47-2.66) (Table 5). Other
EP
predictors included age >65 (OR 1.70, 95% CI 1.50-1.93), sepsis (OR 1.57, 95% CI 1.30-1.90), dialysis (OR 1.50, 95% CI 1.26-1.78), and history of peripheral vascular disease (OR 1.49, 95% CI
AC C
1.14-1.95). Additional lab values associated with 30-day mortality included pre-operative creatinine greater than or equal to 2 (OR 1.74, 95% CI 1.52-1.98) and pre-operative platelets < 100 (OR 1.72, 95% CI 1.48-2.00).
In summary, 44% of exploratory laparotomies on ventilated patients result in death. 31% of exploratory laparotomies were isolated laparotomies meaning that either nothing was found intraoperatively to require further interventions or interventions would not change prognosis.
8
ACCEPTED MANUSCRIPT
Ventilated patients greater than 65, with cancer, or with liver disease were nearly twice as likely to die within 30-days of surgery (p< 0.001, OR 1.91). Regardless of outcome, ventilated patients
RI PT
status post abdominal surgery will spend an average of 26 days in the hospital.
Discussion
This study clearly demonstrates that the mortality of critically ill patients who have been on a
SC
ventilator for 48 hours is high. Procedures involving colon, stomach and small bowel had higher mortality than those involving appendix and gallbladder but overall mortality is 44.2%. Those who
M AN U
underwent exploratory laparotomy alone had a 51.5% 30-day mortality which was higher than some of the laparotomies performed in conjunction with one or two other procedures. Although this initially seemed counterintuitive, we postulate that many of these patients may have had unsalvageable findings such as complete bowel ischemia, which prompted closing the abdomen without performing additional procedures. Unfortunately, due to the nature of the NSQIP
invalidate this hypothesis.
TE D
database, we are not able to obtain further data regarding this subset of patients to either verify or
EP
The multivariate analysis suggests that patients with coma, high protime/INR, esophageal varices, disseminated cancer, platelets less than 100,000/L, creatinine at least 2.0 mg/dL and age greater
AC C
than 65 have the highest risk for mortality. In particular, variables that were associated with liver dysfunction such as esophageal varices, ascites, thrombocytopenia, jaundice and coagulopathy all had significant association with 30-day mortality. Other variables potentially related to liver dysfunction such as coma and renal dysfunction were also significant predictors, but it is difficult to determine if these variables were related to some other non-hepatic disease process. Other studies have validated this, demonstrating increasing mortality with progressive liver dysfunction as measured by Child’s classification and especially in emergency compared to elective surgery in
9
ACCEPTED MANUSCRIPT
cirrhotics [15-18]. Although NSQIP did not delineate which patients were specifically cirrhotic and some of the comorbidities (ascites, thrombocytopenia) are not completely specific to liver function, the significance of jaundice and esophageal varices do indicate that liver dysfunction contributed to
RI PT
mortality.
Previous single-center studies have similarly demonstrated high 30-day mortality in critically ill
SC
patients undergoing abdominal surgery. In a study of 60 septic, critically-ill patients who required laparotomy, Anderson et al. [19] reported a high post-operative mortality rate with only 42%
M AN U
surviving to leave the ICU and 32% surviving to leave the hospital. Ferraris et al. [20] showed that 15 of 29 ICU patients died following abdominal surgery and of those patients over 50 years old, only one survived. While it may have been helpful to directly compare ventilated and nonventilated patients, NSQIP is unable to distinguish elective versus urgent cases and the exact indication for surgery. The individual procedures (cholecystectomies, intestinal resections) have been
TE D
reported in multiple studies to have mortalities of less than 2%, and a few studies have noted increased mortality in emergency procedures compared to elective ones [27-31]. The highest mortalities have been
EP
noted in emergency procedures for toxic megacolon at 16-36% [32-33].
AC C
Despite advances in medical knowledge and care that have occurred since these older studies were conducted, high mortality from abdominal surgery in the ICU continues. These same advances in care have allowed patients to live longer with chronic illnesses and thus require more extensive therapies and costly medical resources in intensive care units to fight complex complications [21]. Recent studies have utilized NSQIP to develop surgical risk calculators to empirically estimate procedure-specific risks [10, 22-24]. The risk calculator developed by Vaid et al. [11] provides a simple tool for predicting operative survival or death using preoperative variables. Bilimoria et al.
10
ACCEPTED MANUSCRIPT
[12] developed a tool to assess both risk of death and the likelihood of post-operative complications. These tools allow patients and physicians to make decisions based upon patientspecific risk factors that predict the expected morbidity and mortality for a particular procedure.
RI PT
However, the calculator may not be practical for daily use on every patient in the ICU as the
variables may change daily and other factors that are not in the calculator, may contribute to risk
SC
and outcome.
Upon detailed examination of the predictors of mortality, it is noted that there were four risk
M AN U
factors that had low odds-ratios for 30-day mortality: BMI>30 (0.78), wound infection (0.78), alcohol history (0.66), and quadriplegia (0.36). Although the exact reason that wound infections were associated with lower mortality is not clear, there are some possible contributing factors. Severe intra-abdominal catastrophe may have prompted surgeons to leave the wound open. Furthermore, with nearly half of the study population dying within the first week, perhaps patients
TE D
did not survive long enough to develop a wound infection. Our study also showed that obese patients with BMI >30 had a decreased mortality. While not initially expected, in retrospect, those patients with BMI <30 would have also included those patients with very low BMI, who were
EP
severely malnourished or sarcopenic. Previous studies have shown that weight loss greater than 10% within 6 months and sarcopenia were associated with post-operative death [25, 26]. Thus our
AC C
grouping of BMI >30 vs BMI <30 may have inadvertently skewed conclusions with respect to mortality. The sample size of quadriplegics (1.6%) and alcoholics (6.1%) who underwent urgent abdominal surgery was extremely small and likely limited accurate conclusions about mortality.
This study did have limitations, as NSQIP is an administrative database and can only be as precise as the trained staff who enter the data and codes. Furthermore, the NSQIP data does rely on some subjective variables that may depend on how meticulous the physicians were at documentation and
11
ACCEPTED MANUSCRIPT
how thoroughly the coders perused the chart. This may vary between institutions and NSQIP staff. More recently NSQIP has added in diagnosis codes but in earlier years, it did not provide a specific diagnosis consistently. As a result, we are unable to determine the exact disease process and
RI PT
reason for each procedure. NSQIP is also unable to give us the details and nuances of the surgical procedure in terms of the extent of intestinal resection, difficulty of the dissection, aberrant
anatomy, previous surgery, or pathology of the resected organ (malignant vs. benign). Difficulty of
SC
the surgery can only be conjectured by operative time and blood transfusions. While NSQIP does account for the presence and level of surgical resident in the case, it does not account for the
M AN U
experience level of the primary surgeon performing the case or the physician managing the patient post-operatively. More specifically in this study, we may have underestimated the problem of high mortality in this population, as we excluded patients who were intubated for less than 48 hours. We also cannot determine if the abdominal surgery was the primary reason for admission or if some of these patients were intubated for another chronic illness and secondarily developed an
TE D
intra-abdominal problem for which surgery was performed. Lastly, in our consideration of common emergency abdominal procedures, non-resectional closures of perforated ulcers were not included in the CPT codes we used to search the database. Thus, there may have been a small number of
EP
critically ill patients undergoing abdominal procedures who were not included in this study.
AC C
In spite of these limitations, this study analyzed a large number of critically ill patients who underwent abdominal surgery after being on the ventilator for at least 48 hours and 44% of them died shortly after the procedure. Those who did survive the operation were hospitalized for an average of 18 additional days. Our study also demonstrated that patients who were in a coma or had disseminated cancer or complications from chronic liver disease had especially poor outcomes. Surgeons and critical care physicians should be aware of the high mortality of abdominal surgery in the critically ill as they approach patients and their families to make decisions about surgical
12
ACCEPTED MANUSCRIPT
interventions. While specific values from risk calculators may help individual patients, defining broad categories of patients with known poor prognosis may potentially assist critical care staff in developing protocols for approaching patients and families with such conditions. Improved
RI PT
education of patients and family members regarding the diagnosis and expected outcome of a
procedure will hopefully result in realistic patient care goals, and when appropriate, the redirection
SC
of resources to palliative care.
Conclusion
M AN U
Critically ill patients who require ventilator support and undergo abdominal surgery have high mortality, with nearly half of these patients dying in the first week. Regardless of outcome, patients will spend a mean of 26 total days in the hospital, at risk for nosocomial infection and other potential hospital-associated complications. Those patients who have esophageal varices, pre-op INR greater than 3.0, are comatose, have disseminated cancer, or are older than 65 years, are at the
TE D
highest risk. With this knowledge of particularly high mortality in ventilated ICU patients with cancer, liver disease or advanced age, surgeons should have thorough discussions with the patients
AC C
EP
and family members to consider redirecting efforts to palliative care in these patients.
13
ACCEPTED MANUSCRIPT
Acknowledgments The American College of Surgeons National Surgical Quality Improvement Program and the hospitals participating in the ACS NSQIP are the source of the data used herein; they have not
RI PT
verified and are not responsible for the statistical validity of the data analysis or the conclusions
AC C
EP
TE D
M AN U
SC
derived by the authors.
14
ACCEPTED MANUSCRIPT
References 1. Angus DC, Barnato AE, Linde-Zwirble WT, et al. Robert Wood Johnson Foundation ICU EndOf-Life Peer Group. Use of intensive care at the end of life in the United States: an
RI PT
epidemiologic study. Crit Care Med, 2004;32(3):638-43. 2. Halpern NA, Pastores SM, Greenstein RJ: Critical care medicine in the United States 1985–2000: An analysis of bed numbers, use, and costs. Crit Care Med, 2004;32:1254–59.
SC
3. Lubitz JD, Riley GF. Trends in Medicare payments in the last year of life. N Engl J Med, 1993;328(15):1092-6.
Serv Res, 2010;45(2):565-76.
M AN U
4. Riley GF, Lubitz JD. Long-term trends in Medicare payments in the last year of life. Health
5. Centers for Medicare & Medicaid Services. National Health Expenditures 2013 Highlights. Available at: http://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trendsand-Reports/NationalHealthExpendData/Downloads/highlights.pdf Unknown original
TE D
publication date. Accessed January 11, 2015.
6. Dasta JF, McLaughlin TP, Mody SH, Piech CT. Daily cost of an intensive care unit day: the contribution of mechanical ventilation. Critical Care Med, 2005;33(6):1266-71.
EP
7. Huynh T, Kleerup E, Wiley J, et al. The Frequency and Cost of Treatment Perceived to Be Futile in Critical Care. JAMA Internal Medicine, 2013;173(20):1887-94.
AC C
8. Teno JM, Murphy D, Lynn J, et al. Prognosis-based futility guidelines: dose anyone win? SUPPORT Investigators. Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatment. J Am Geriatr Soc, 1994;42(11):1202-7.
9. Teno JM, Lynn J, Connors AF Jr, et al. The illusion of end of life resource savings with advance directives.
SUPPORT Investigators.
Study to Understand Prognoses and
Preferences for Outcomes and Risks of Treatment. J Am Geriatr Soc, 1997;45(4):513-8.
15
ACCEPTED MANUSCRIPT
10. National Surgical Quality Improvement Program. Chicago, IL: American College of Surgeons; 2002-2015. Available at: www.acsnsqip.org. Accessed January 31, 2015. 11. Vaid S, Bell T, Grim R, Ahuja V. Predicting risk of death in general surgery patients on the
RI PT
basis of preoperative variables using American College of Surgeons National Surgical Quality Improvement Program data. Perm J, 2012;16(4):10-7.
12. Bilimoria K, Liu Y, Paruch J, et al. Development and Evaluation of the Universal ACS NSQIP
Surgeons. J Am Coll Surg, 2013;217(5):833-42.
SC
Surgical Risk Calculator: A Decision Aid and Informed Consent Tool for Patients and
M AN U
13. Cohen ME, Bilimoria KR, Ko CY, et al. Development of an American College of Surgeons National Surgery Quality Improvement Program: morbidity and mortality risk calculator for colorectal surgery. J Am Coll Surg, 2009;208(6):1009-16.
14. American Medical Association. CPT 2014 Professional Edition. American Medical Association Press, 2013.
TE D
15. Garrison RN, Cryer HM, Howard DA, Polk HC. Clarification of risk factors for abdominal operations in patients with hepatic cirrhosis. Ann Surg,1984;199(6):648-55. 16. Farnsworth N, Fagan SP, Berger DH, Awad SS. Child-Turcotte-Pugh versus MELD score as a
EP
predictor of outcome after elective and emergent surgery in cirrhotic patients. Am J Surg. 2004;188(5):580-3.
AC C
17. Franzetta M, Raimondo D, Giammanco M, et al. Prognostic factors of cirrhotic patients in extra-hepatic surgery. Minerva Chir,2003;58(4):541-4.
18. Mansour A, Watson W, Shayani V, et al. Abdominal operations in patients with cirrhosis: still a major surgical challenge. Surgery, 1997;122:730–5.
19. Anderson I, Fearon K, Grant I. Laparotomy for abdominal sepsis in the critically ill. Br J Surg, 1996;83(4):535-9.
16
ACCEPTED MANUSCRIPT
20. Ferraris V. Exploratory Laparotomy for Potential Abdominal Sepsis in Patients with Multiple Organ Failure. Arch Surg, 1983;118(10):1130-3. 21. American Hospital Association. Are Medicare Patients Getting Sicker? Trendwatch.
RI PT
December 2012.
22. Tzeng CW, Cooper AB, Vauthey JN, et al. Predictors of morbidity and mortality after hepatectomy in elderly patients: analysis of 7621 NSQIP patients. HPB
SC
(Oxford), 2014;16(5):459-68.
23. Kneuertz PJ, Pitt HA, Bilimoria KY, et al. Risk of morbidity and mortality following hepto-
M AN U
pancreato-biliary surgery. J Gastrointest Surg, 2012;16(9):1727-35.
24. Greenblatt DY, Kelly KJ, Rajamanickam V, et al. Preoperative factors predict perioperative morbidity and mortality after pancreaticoduodenectomy. Ann Surg Oncol, 2011;18(8):2126-35.
25. Englesbe MJ, Patel SP, He K. et al. Sarcopenia and Mortality after Liver Transplantation. J
TE D
Am Coll Surg, 2010;211(2):271-278.
26. Slim K, Panis Y, Alves A, et al. Predicting Postoperative Mortality in Patients Undergoing Colorectal Surgery. World J Surg, 2006;30:100-106.
EP
27. Steiner CA, Bass EB, Talamini MA, et al. Surgical rates and operative mortality for open and LC in Maryland. N Engl J Med 1994;330:403.
AC C
28. Bittner R, Butters M, Ulrich M, et al. Total Gastrectomy. Updated operative mortality and long-term survival with particular reference to patients older than 70 years of age. Ann Surg 1996; 224(1):37-42.
29. Antoniou SA, Antoniou GA, Koch OO, et al. Meta-analysis of laparoscopic versus open repair of perforated peptic ulcer. JSLS 2013;17(1):15-22.
17
ACCEPTED MANUSCRIPT
30. Juo Y, Hyder O, Haider A H, et al. Is Minimally Invasive Colon Resection Better than Traditional Approaches? First Comprehensive National Examination With Propensity Score Matching. JAMA Surg 2014;149(2):177-84.
RI PT
31. Tottrup A, Erichsen R, Svaerke C, et al. Thirty-day mortality after elective and emergency total colectomy in Danish patients with inflammatory bowel disease: a population-based nationwide cohort study. BMJ Open 2012;2(2):e000823.
SC
32. Tapani MJ, Olavi KH. Surgical Management of Toxic Megacolon. Hepatogastroenterology 2014;61(131):638-41.
M AN U
33. Hall JF, Berger D. Outcome of colectomy for Clostridium difficile colitis: a plea for early
AC C
EP
TE D
surgical management. Am J Surg 2008;196(3):384-8.
18
ACCEPTED MANUSCRIPT
Table 1. Patient Characteristics
AC C
EP
TE D
M AN U
SC
RI PT
Patients Hypertension requiring medications 3154 (64.4%) Previous septic shock 3026 (61.7%) Bleeding disorder 1435 (29.3%) Diabetes 1291 (26.3%) History of smoking 1300 (26.5%) History of COPD 984 (20.1%) Concurrent pneumonia 959 (19.6%) Ascites 885 (18.1%) Renal failure 879 (17.9%) Dialysis 784 (16%) Previous SIRS 725 (14.8%) CVA 657 (13.4%) Congestive heart failure 652 (13.3%) Previous major cardiac surgery 642 (13.1%) Steroid 596 (12.2%) Previous sepsis 513 (10.5%) Previous percutaneous coronary intervention 416 (8.5%) MI within 6 months 356 (7.3%) Weight loss (>10% within 6 months) 313 (6.4%) History of alcohol use 301 (6.1%) Peripheral vascular disease 262 (5.3%) Coma 226 (4.6%) Disseminated cancer 214 (4.4%) History of angina 153 (3.1%) Prior chemotherapy 135 (2.8%) Quadriplegia 80 (1.6%) Paraplegia 74 (1.5%) Prior radiation 58 (1.2%) Esophageal varices 44 (0.9%) COPD = Chronic Obstructive Pulmonary Disease. MI = Myocardial Infarction. CVA = Cerebrovascular accident. SIRS = Systemic Inflammatory Response Syndrome
19
ACCEPTED MANUSCRIPT
Table 2. Procedure specific 30-day mortality
51.3% 45.2% 42.2% 28.7% 27.35% 12.5% 51.5% 39.8% 50.0% 54.2%
SC
1570 (32.0%) 1022 (20.9%) 1881 (38.4%) 87 (1.8%) 245 (5.0%) 96 (2.0%) 1521 (31.0%) 3032 (61.9%) 324 (6.6%) 24 (0.5%)
M AN U
Primary procedure Exploratory laparotomy Small bowel Colon Gastric Cholecystectomy Appendectomy Exploratory laparotomy alone One abdominal procedure * Two abdominal procedures* Three or more abdominal procedures*
30-Day Mortality
RI PT
#Patients (% total)
AC C
EP
TE D
*Refers to procedure exclusive of coding for exploratory laparotomy
20
AC C
EP
TE D
M AN U
SC
Table 3. Complications Complication Frequency Failure to wean from ventilator within 48-hours 57.4% Required transfusion 26.6% Pneumonia 16.9% Septic shock 14.2% Re-intubation 11% Acute renal failure 10.4% Sepsis 9.8% Cardiac arrest 7.9% Superficial wound infection 3.5% Wound dehiscence 3.4% Coma 2.3 Deep wound infection 2.1% MI 2% CVA 1.9% Pulmonary emboli 1.2% CVA = Cerebrovascular accident. MI = Myocardial Infarction.
RI PT
ACCEPTED MANUSCRIPT
21
ACCEPTED MANUSCRIPT
Table 4. Outcome measures Average (SD) 25.8 (27.6) days 6.8 (7.9) days 103.7 (68.3) minutes 1.2 (2.5) units 18.8 (21.1) days
AC C
EP
TE D
M AN U
SC
RI PT
Total hospital length of stay Days operation to death Operation time Blood transfusions Days operation to discharge
22
ACCEPTED MANUSCRIPT
Table 5. Significant risk factors predictive of 30-day mortality P-value
Coma
2.63 (1.95-3.55)
P<0.001
Pre-op INR >3.0
2.60 (1.79-3.79)
P<0.001
Esophageal varices
2.44 (1.21-4.91)
P=0.002
Disseminated Cancer
1.98 (1.47-2.66)
P<0.001
Pre-op Creatinine ≥2 mg/dL
1.74 (1.52-1.98)
P<0.001
Pre-op Platelets < 100 (x109/L)
1.72 (1.48-2.00)
Age >65 years
1.70 (1.50-1.93)
Any Sepsis/SIRS
1.57 (1.30-1.90)
Dialysis
1.50 (1.26-1.78)
P<0.001
History of PVD
1.49 (1.14-1.95)
P=0.004
Ascites
1.29 (1.10-1.51)
P=0.002
Pre-op Bilirubin ≥2mg/dL
1.29 (1.10-1.50)
P=0.002
History of COPD
1.23 (1.06-1.44)
P=0.007
Alcohol history
TE D
Wound Infection BMI > 30
SC
RI PT
Odds-ratio (95% Confidence Interval)
P<0.001 P<0.001 P<0.001
M AN U
Risk Factor
0.80 (0.69-0.93)
P=0.003
0.76 (0.69-0.87)
P<0.001
0.73 (0.57-0.96)
P=0.002
AC C
EP
Quadriplegia 0.55 (0.31-0.98) P=0.038 INR = International Normalized Ration. SIRS = Systemic Inflammatory Response Syndrome. PVD = Peripheral Vascular Disease. Na = Serum Sodium. COPD = Chronic Obstructive Pulmonary Disease. BMI = Body Mass Index.
23
S0002-9610(16)30197-0
DOI:
10.1016/j.amjsurg.2015.12.029
Reference:
AJS 11891
To appear in:
The American Journal of Surgery
Received Date: 29 July 2015 Revised Date:
5 December 2015
Accepted Date: 10 December 2015
Please cite this article as: Lu N, Marumoto A, Wong LL, Outcomes of abdominal surgery in patients receiving mechanical ventilation for more than 48 hours, The American Journal of Surgery (2016), doi: 10.1016/j.amjsurg.2015.12.029. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Abstract Background Abdominal surgery in critically-ill patients has high mortality; contributing to high U.S.
RI PT
healthcare costs. This study sought to identify specific predictors of mortality in this population.
Methods Using the National Surgical Quality Improvement Program database 2006–2012, we
SC
identified 4901 patients who were intubated for more than 48 hours prior to undergoing common abdominal procedures. Mortality and predictors of mortality were determined using chi-
M AN U
square/regression analysis.
Results Overall 30-day mortality was 44.2% with increasing mortality for additional procedures performed. Ventilated patients with the following pre-operative risk factors were two to three
TE D
times as likely to die within 30 days of surgery: age greater than 65 years-old, coma, pre-op INR > 3.0, esophageal varices, and disseminated cancer.
EP
Conclusions Mortality is significant in ventilated patients who undergo abdominal surgery and is especially high with advanced age, disseminated cancer, and complications of liver disease.
AC C
Physicians should carefully discuss this with patients/family and consider palliative options when appropriate.
ACCEPTED MANUSCRIPT
Keywords Abdominal surgery Critical care
RI PT
Acute abdomen NSQIP
AC C
EP
TE D
M AN U
SC
Mortality
ACCEPTED MANUSCRIPT
Summary for the table of contents Abdominal surgery in critically-ill patients has high mortality; contributing to high U.S. healthcare costs. Using the National Surgical Quality Improvement Program database 2006–2012, we
RI PT
identified 4901 patients who were intubated for more than 48 hours prior to undergoing common abdominal procedures. Overall 30-day mortality was 44.2% with increasing mortality for additional procedures performed. Mortality is significant in ventilated patients who undergo abdominal
SC
surgery and is especially high with advanced age, disseminated cancer, and complications of liver
AC C
EP
TE D
M AN U
disease
ACCEPTED MANUSCRIPT
Outcomes of abdominal surgery in patients receiving mechanical ventilation for more than
RI PT
48 hours
Ning Lu, M.D.a
SC
Email: [email protected] Ashley Marumoto, M.D.a
Linda L. Wong, M.D.a, b
aDepartment
of Surgery, University of Hawaii, John A. Burns School of Medicine,
Honolulu, Hawaii 96813
TE D
1356 Lusitana Street, Sixth Floor
b Cancer
M AN U
Email: [email protected]
Center, University of Hawaii
701 Ilalo Street, Suite 600
EP
Honolulu, Hawai'i 96813
AC C
Corresponding author:
Linda L. Wong MD
550 South Beretania St, Suite 403 Honolulu, Hawaii 96813 Phone: (808)-523-5033 FAX: (808)-528-4940 E mail: [email protected]
1
ACCEPTED MANUSCRIPT
Introduction Critical care is expensive and has high mortality with 20% of all deaths in the United States occurring during or shortly after an intensive care unit (ICU) admission [1]. In 2000, critical care
RI PT
medicine accounted for 13.3% of hospital costs and 0.56% of the gross national domestic product in the United States [2]. Of the $585.7 billion Medicare spent in 2013, 25-30% was used towards the 5% of Medicare beneficiaries who died, which was six times the cost of Medicare spending for a
SC
survivor [3-5]. The first two days of intensive care ICU admission accounted for the highest costs, with mechanical ventilation resulting in higher daily costs and increasing the cost 2.5 fold
M AN U
compared to non-ventilated patients [6].
Admission to the ICU may be considered a therapeutic trial, whereby initial aggressive measures are transitioned to palliative once it is clear that meaningful outcomes cannot be achieved [7]. Although there is no formal definition for “futile care”, it is generally understood as care that cannot
TE D
achieve a patient’s quality of life goals. Futile care is often scrutinized for utilizing much of healthcare resources with little return. In 1995, The Study to Understand Prognosis and Preferences for Outcomes and Risk of Treatment (SUPPORT) [8,9] attempted to identify patients
EP
with poor survival prognoses and intervene so as to reduce futile care and interventions in these patients. Despite the availability of prognostic information and physician efforts to communicate
AC C
this to patients and their families, there was no reduction of futile care [8,9]. Although there is no objective means of prospectively identifying patients whose care will be futile, physicians should consider palliative care options and be sensitive to cost and utilization of resources.
Abdominal surgery is particularly risky in critically ill patients. These operations may contribute to the length of ICU admission, length of hospital stay and high cost of healthcare, but may have limited impact on survival or quality of life. A few studies have addressed mortality following
3
ACCEPTED MANUSCRIPT
exploratory laparotomy in critically ill patients but these were small samples, single center and before the era of evidence-based medicine in large databases such as the American College of Surgeons’ National Surgical Quality Improvement Program (NSQIP) [10]. Surgical risk calculators
RI PT
have been developed and are useful in estimating an individual mortality based on procedure and risk factors [11-13]. This study attempts to provide evidence-based data that can be more broadly generalized on mortality rates and predictors of mortality in critically ill patients compared to prior
SC
single-center studies. It also hopes to supplement the current NSQIP calculator by providing additional data on pre-operative risk factors and post-operative mortality that will provide
M AN U
physicians with prognostic values that can be utilized on a daily basis to help patients and their families make difficult management decisions regarding their care in the ICU.
Materials and Methods
This study is a retrospective review of the prospectively collected NSQIP database from 2006
TE D
through 2012. The database is compiled from 373 hospitals nationwide and provides information on 30-day, risk-adjusted surgical outcomes, which allows hospitals across the nation to compare outcomes. Pre-operative through 30-day post-operative variables are collected on randomly
EP
assigned patients, including patient demographics, surgical profile, preoperative risk assessment, laboratory values, operative information, and 30-day morbidity and mortality rates. A highly
AC C
trained Surgical Clinical Reviewer collects the data. All reviewers receive extensive initial training prior to starting data collection and ongoing training via continuing education. NSQIP monitors accrual rates and data sampling methodologies and conducts audits on a random basis, ensuring highly reliable data.
Patients: Common Procedural Terminology (CPT) codes [14] for frequently performed emergent abdominal procedures were used to identify patients from the NSQIP database: 49000, 43632,
4
ACCEPTED MANUSCRIPT
43635, 43621, 47600, 47562, 44120, 44125, 44950, 44960, 44970, 44143, 44320, 44140, 44150, 44204. During the time period from January 1, 2006 thru December 31, 2012, NSQIP collected data on 2,320,898 patients. Increasing numbers of patients were identified annually with 152,468
RI PT
patients recorded in 2006, the year of NSQIP’s inception and 543,885 patients added in 2012. Note that the 2006 NSQIP database did include a small number of cases that underwent a surgical
procedure in 2005 but data from the entirety of 2005 is not represented in this file. Because NSQIP
SC
does not identify whether patients are located in an ICU or on a general ward at the time of surgery, we used the parameter of “ventilated in the last 48 hours” as a surrogate for being located in an
M AN U
intensive care unit. We thus identified 17,679 (0.7%) patients who were on the ventilator for at least 48 hours and 4,901 of these patients underwent abdominal surgery as defined by the CPT codes. This was the cohort used in the study and our definition of a critically ill patient undergoing abdominal surgery. This analysis would potentially exclude patients who were critically ill but not intubated and those who were intubated for less than 48 hours prior to surgery. Trauma patients
TE D
are also excluded from the NSQIP database. For this study, we identified patients who had one of the primary CPT codes for abdominal surgery and categorized them by organ system: cholecystectomy, appendectomy, gastric, colon or small bowel procedures. We then looked at the
EP
subsequent 3 additional CPT codes to determine if other abdominal procedures were performed at this surgery. Patients were categorized as to having received exploratory laparotomy alone, one
AC C
abdominal procedure, two abdominal procedures, or three or more abdominal procedures.
NSQIP data [10] included demographics: age, sex, ethnicity, height, weight, and BMI. Comorbidities included diabetes, smoking, alcohol use, chronic obstructive pulmonary disease (COPD), concurrent pneumonia, ascites, congestive heart failure (CHF), esophageal varices, myocardial infarction within 6 months, previous percutaneous coronary intervention, angina, hypertension requiring medications, peripheral vascular disease, renal failure, dialysis, coma, cerebrovascular accident,
5
ACCEPTED MANUSCRIPT
paralysis, wound infection, disseminated cancer, steroid use, significant weight loss (10%) in the previous 6 months, bleeding disorder, prior chemotherapy, radiation, previous sepsis, systemic
RI PT
inflammatory response syndrome (SIRS), or septic shock.
Pre-operative laboratory values were divided into categorical variables as defined by: abnormal serum sodium (less than or equal to 130 or greater than or equal to 150 mmol/L) , serum
SC
creatinine greater than or equal to 2.0 mg/dL, albumin less than or equal to 3.0 gm/dL, serum bilirubin greater than or equal to 2.0 mg/dL, AST greater than 200 IU/L, alkaline phosphatase
M AN U
greater than 150 IU/L, abnormal WBC (less than 3 or greater than 12 x 109/L), hematocrit less than 30%, platelet count less than 100 x 109/L, and INR greater than 3.
Post-operative complications included superficial or deep wound infection, wound dehiscence, pneumonia, re-intubation, pulmonary embolism, failure to wean from ventilator in 48 hours, renal
TE D
insufficiency, acute renal failure, cerebrovascular accident, coma, cardiac arrest, myocardial infarction, transfusions, sepsis, septic shock and return to the operating room. Outcome variables included time in the operating room, number of units of blood transfused, return to the operating
EP
room, hospital length of stay, and 30-day mortality.
AC C
Data Endpoints: The goals of the study were to identify pre-operative risk factors for intraoperative and post-operative mortality in critically ill patients who underwent emergency abdominal surgery. NSQIP provided information on length of stay and days from operation to death. This information was used to identify patients who had died within 30-days of their procedure.
6
ACCEPTED MANUSCRIPT
Data Analysis: All analyses were performed using SPSS statistical software. Demographics (age greater than 65 years, gender), comorbidities and preoperative laboratory studies based on the categorical variables above were compared between 30-day survivors and 30-day non-survivors
RI PT
using the chi-square test. With this univariate analysis, all variables in which p <0.10 were entered into multivariate analysis to identify significant variables at p<0.05. Logistic regression was then used to calculate odds-ratios (OR) and 95% confidence intervals (CI) to identify the best predictors
SC
of 30-day mortality.
M AN U
Results
In the overall cohort of 4,901 critically ill patients who underwent emergency abdominal surgery, 53% were male, mean age was 63.9 years, and 2,560 (52.2%) were over the age of 65 years. Other characteristics are in Table 1. Preoperative laboratory studies by category included: abnormal sodium (9.6%), creatinine greater than or equal to 2.0 mg/dL (42.8%), albumin less than or equal
TE D
to 3.0 gm/dL (73.4%), bilirubin greater than or equal to 2.0 mg/dL (21.1%), AST more than 200 IU/L (9%) , alkaline phosphatase greater than 150 IU/L (16.2%), abnormal WBC (55.6%),
EP
hematocrit less than 30% (42.5%) , and platelet count less than 100 x 109/L (22.0%).
The most common procedures were those relating to the colon (1881 patients, 38.4%) followed by
AC C
exploratory laparotomy (1570 patients, 32.0%). Other primary procedures were distributed by type (Table 2). Of the 1570 patients who had exploratory laparotomy listed as the primary procedure, 1521 patients did not undergo any other abdominal procedure while 49 patients had an additional abdominal procedure performed. The 1521 patients that had only an exploratory laparotomy had a 30-day post-operative mortality rate at 51.5%. For those who were coded for 1, 2, or 3 additional abdominal procedures, the mortality rate was 39.8%, 50% and 54.2% respectively (Table 2).
7
ACCEPTED MANUSCRIPT
The most common complications included failure to wean from the ventilator within 48 hours (57.4%), transfusion requirement (26.6%), pneumonia (16.9%), septic shock (14.2%), re-
RI PT
intubation (11%), and acute renal failure (10.4%). Other complications are listed in table 3.
The major outcome measures are as follows (Table 4): In this cohort, 26.6% patients required
SC
blood transfusion and mean number of transfusions was 1.2 units (SD 2.5 units). Mean operative time was 103.7 minutes (SD 68.3 minutes) and 28% required a return to the operating room after
M AN U
the initial operation. Mean hospital length of stay was 25.8 days (SD 27.7 days). Of the 2165 patients who died, the time from operation to death was a mean of 6.75 days (range 0-30 days, SD 7.92 days). Of the survivors mean hospital length of stay following the procedure was 18.8 days.
TE D
The pre-operative factors best predicting 30-day mortality included: coma (OR 2.63, 95% CI 1.953.55), pre-operative INR greater than 3 (OR 2.6, 95% CI 1.79-3.79), esophageal varices (OR 2.44, 95% CI 1.21-4.91), and disseminated cancer (OR 1.98, 95% CI 1.47-2.66) (Table 5). Other
EP
predictors included age >65 (OR 1.70, 95% CI 1.50-1.93), sepsis (OR 1.57, 95% CI 1.30-1.90), dialysis (OR 1.50, 95% CI 1.26-1.78), and history of peripheral vascular disease (OR 1.49, 95% CI
AC C
1.14-1.95). Additional lab values associated with 30-day mortality included pre-operative creatinine greater than or equal to 2 (OR 1.74, 95% CI 1.52-1.98) and pre-operative platelets < 100 (OR 1.72, 95% CI 1.48-2.00).
In summary, 44% of exploratory laparotomies on ventilated patients result in death. 31% of exploratory laparotomies were isolated laparotomies meaning that either nothing was found intraoperatively to require further interventions or interventions would not change prognosis.
8
ACCEPTED MANUSCRIPT
Ventilated patients greater than 65, with cancer, or with liver disease were nearly twice as likely to die within 30-days of surgery (p< 0.001, OR 1.91). Regardless of outcome, ventilated patients
RI PT
status post abdominal surgery will spend an average of 26 days in the hospital.
Discussion
This study clearly demonstrates that the mortality of critically ill patients who have been on a
SC
ventilator for 48 hours is high. Procedures involving colon, stomach and small bowel had higher mortality than those involving appendix and gallbladder but overall mortality is 44.2%. Those who
M AN U
underwent exploratory laparotomy alone had a 51.5% 30-day mortality which was higher than some of the laparotomies performed in conjunction with one or two other procedures. Although this initially seemed counterintuitive, we postulate that many of these patients may have had unsalvageable findings such as complete bowel ischemia, which prompted closing the abdomen without performing additional procedures. Unfortunately, due to the nature of the NSQIP
invalidate this hypothesis.
TE D
database, we are not able to obtain further data regarding this subset of patients to either verify or
EP
The multivariate analysis suggests that patients with coma, high protime/INR, esophageal varices, disseminated cancer, platelets less than 100,000/L, creatinine at least 2.0 mg/dL and age greater
AC C
than 65 have the highest risk for mortality. In particular, variables that were associated with liver dysfunction such as esophageal varices, ascites, thrombocytopenia, jaundice and coagulopathy all had significant association with 30-day mortality. Other variables potentially related to liver dysfunction such as coma and renal dysfunction were also significant predictors, but it is difficult to determine if these variables were related to some other non-hepatic disease process. Other studies have validated this, demonstrating increasing mortality with progressive liver dysfunction as measured by Child’s classification and especially in emergency compared to elective surgery in
9
ACCEPTED MANUSCRIPT
cirrhotics [15-18]. Although NSQIP did not delineate which patients were specifically cirrhotic and some of the comorbidities (ascites, thrombocytopenia) are not completely specific to liver function, the significance of jaundice and esophageal varices do indicate that liver dysfunction contributed to
RI PT
mortality.
Previous single-center studies have similarly demonstrated high 30-day mortality in critically ill
SC
patients undergoing abdominal surgery. In a study of 60 septic, critically-ill patients who required laparotomy, Anderson et al. [19] reported a high post-operative mortality rate with only 42%
M AN U
surviving to leave the ICU and 32% surviving to leave the hospital. Ferraris et al. [20] showed that 15 of 29 ICU patients died following abdominal surgery and of those patients over 50 years old, only one survived. While it may have been helpful to directly compare ventilated and nonventilated patients, NSQIP is unable to distinguish elective versus urgent cases and the exact indication for surgery. The individual procedures (cholecystectomies, intestinal resections) have been
TE D
reported in multiple studies to have mortalities of less than 2%, and a few studies have noted increased mortality in emergency procedures compared to elective ones [27-31]. The highest mortalities have been
EP
noted in emergency procedures for toxic megacolon at 16-36% [32-33].
AC C
Despite advances in medical knowledge and care that have occurred since these older studies were conducted, high mortality from abdominal surgery in the ICU continues. These same advances in care have allowed patients to live longer with chronic illnesses and thus require more extensive therapies and costly medical resources in intensive care units to fight complex complications [21]. Recent studies have utilized NSQIP to develop surgical risk calculators to empirically estimate procedure-specific risks [10, 22-24]. The risk calculator developed by Vaid et al. [11] provides a simple tool for predicting operative survival or death using preoperative variables. Bilimoria et al.
10
ACCEPTED MANUSCRIPT
[12] developed a tool to assess both risk of death and the likelihood of post-operative complications. These tools allow patients and physicians to make decisions based upon patientspecific risk factors that predict the expected morbidity and mortality for a particular procedure.
RI PT
However, the calculator may not be practical for daily use on every patient in the ICU as the
variables may change daily and other factors that are not in the calculator, may contribute to risk
SC
and outcome.
Upon detailed examination of the predictors of mortality, it is noted that there were four risk
M AN U
factors that had low odds-ratios for 30-day mortality: BMI>30 (0.78), wound infection (0.78), alcohol history (0.66), and quadriplegia (0.36). Although the exact reason that wound infections were associated with lower mortality is not clear, there are some possible contributing factors. Severe intra-abdominal catastrophe may have prompted surgeons to leave the wound open. Furthermore, with nearly half of the study population dying within the first week, perhaps patients
TE D
did not survive long enough to develop a wound infection. Our study also showed that obese patients with BMI >30 had a decreased mortality. While not initially expected, in retrospect, those patients with BMI <30 would have also included those patients with very low BMI, who were
EP
severely malnourished or sarcopenic. Previous studies have shown that weight loss greater than 10% within 6 months and sarcopenia were associated with post-operative death [25, 26]. Thus our
AC C
grouping of BMI >30 vs BMI <30 may have inadvertently skewed conclusions with respect to mortality. The sample size of quadriplegics (1.6%) and alcoholics (6.1%) who underwent urgent abdominal surgery was extremely small and likely limited accurate conclusions about mortality.
This study did have limitations, as NSQIP is an administrative database and can only be as precise as the trained staff who enter the data and codes. Furthermore, the NSQIP data does rely on some subjective variables that may depend on how meticulous the physicians were at documentation and
11
ACCEPTED MANUSCRIPT
how thoroughly the coders perused the chart. This may vary between institutions and NSQIP staff. More recently NSQIP has added in diagnosis codes but in earlier years, it did not provide a specific diagnosis consistently. As a result, we are unable to determine the exact disease process and
RI PT
reason for each procedure. NSQIP is also unable to give us the details and nuances of the surgical procedure in terms of the extent of intestinal resection, difficulty of the dissection, aberrant
anatomy, previous surgery, or pathology of the resected organ (malignant vs. benign). Difficulty of
SC
the surgery can only be conjectured by operative time and blood transfusions. While NSQIP does account for the presence and level of surgical resident in the case, it does not account for the
M AN U
experience level of the primary surgeon performing the case or the physician managing the patient post-operatively. More specifically in this study, we may have underestimated the problem of high mortality in this population, as we excluded patients who were intubated for less than 48 hours. We also cannot determine if the abdominal surgery was the primary reason for admission or if some of these patients were intubated for another chronic illness and secondarily developed an
TE D
intra-abdominal problem for which surgery was performed. Lastly, in our consideration of common emergency abdominal procedures, non-resectional closures of perforated ulcers were not included in the CPT codes we used to search the database. Thus, there may have been a small number of
EP
critically ill patients undergoing abdominal procedures who were not included in this study.
AC C
In spite of these limitations, this study analyzed a large number of critically ill patients who underwent abdominal surgery after being on the ventilator for at least 48 hours and 44% of them died shortly after the procedure. Those who did survive the operation were hospitalized for an average of 18 additional days. Our study also demonstrated that patients who were in a coma or had disseminated cancer or complications from chronic liver disease had especially poor outcomes. Surgeons and critical care physicians should be aware of the high mortality of abdominal surgery in the critically ill as they approach patients and their families to make decisions about surgical
12
ACCEPTED MANUSCRIPT
interventions. While specific values from risk calculators may help individual patients, defining broad categories of patients with known poor prognosis may potentially assist critical care staff in developing protocols for approaching patients and families with such conditions. Improved
RI PT
education of patients and family members regarding the diagnosis and expected outcome of a
procedure will hopefully result in realistic patient care goals, and when appropriate, the redirection
SC
of resources to palliative care.
Conclusion
M AN U
Critically ill patients who require ventilator support and undergo abdominal surgery have high mortality, with nearly half of these patients dying in the first week. Regardless of outcome, patients will spend a mean of 26 total days in the hospital, at risk for nosocomial infection and other potential hospital-associated complications. Those patients who have esophageal varices, pre-op INR greater than 3.0, are comatose, have disseminated cancer, or are older than 65 years, are at the
TE D
highest risk. With this knowledge of particularly high mortality in ventilated ICU patients with cancer, liver disease or advanced age, surgeons should have thorough discussions with the patients
AC C
EP
and family members to consider redirecting efforts to palliative care in these patients.
13
ACCEPTED MANUSCRIPT
Acknowledgments The American College of Surgeons National Surgical Quality Improvement Program and the hospitals participating in the ACS NSQIP are the source of the data used herein; they have not
RI PT
verified and are not responsible for the statistical validity of the data analysis or the conclusions
AC C
EP
TE D
M AN U
SC
derived by the authors.
14
ACCEPTED MANUSCRIPT
References 1. Angus DC, Barnato AE, Linde-Zwirble WT, et al. Robert Wood Johnson Foundation ICU EndOf-Life Peer Group. Use of intensive care at the end of life in the United States: an
RI PT
epidemiologic study. Crit Care Med, 2004;32(3):638-43. 2. Halpern NA, Pastores SM, Greenstein RJ: Critical care medicine in the United States 1985–2000: An analysis of bed numbers, use, and costs. Crit Care Med, 2004;32:1254–59.
SC
3. Lubitz JD, Riley GF. Trends in Medicare payments in the last year of life. N Engl J Med, 1993;328(15):1092-6.
Serv Res, 2010;45(2):565-76.
M AN U
4. Riley GF, Lubitz JD. Long-term trends in Medicare payments in the last year of life. Health
5. Centers for Medicare & Medicaid Services. National Health Expenditures 2013 Highlights. Available at: http://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trendsand-Reports/NationalHealthExpendData/Downloads/highlights.pdf Unknown original
TE D
publication date. Accessed January 11, 2015.
6. Dasta JF, McLaughlin TP, Mody SH, Piech CT. Daily cost of an intensive care unit day: the contribution of mechanical ventilation. Critical Care Med, 2005;33(6):1266-71.
EP
7. Huynh T, Kleerup E, Wiley J, et al. The Frequency and Cost of Treatment Perceived to Be Futile in Critical Care. JAMA Internal Medicine, 2013;173(20):1887-94.
AC C
8. Teno JM, Murphy D, Lynn J, et al. Prognosis-based futility guidelines: dose anyone win? SUPPORT Investigators. Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatment. J Am Geriatr Soc, 1994;42(11):1202-7.
9. Teno JM, Lynn J, Connors AF Jr, et al. The illusion of end of life resource savings with advance directives.
SUPPORT Investigators.
Study to Understand Prognoses and
Preferences for Outcomes and Risks of Treatment. J Am Geriatr Soc, 1997;45(4):513-8.
15
ACCEPTED MANUSCRIPT
10. National Surgical Quality Improvement Program. Chicago, IL: American College of Surgeons; 2002-2015. Available at: www.acsnsqip.org. Accessed January 31, 2015. 11. Vaid S, Bell T, Grim R, Ahuja V. Predicting risk of death in general surgery patients on the
RI PT
basis of preoperative variables using American College of Surgeons National Surgical Quality Improvement Program data. Perm J, 2012;16(4):10-7.
12. Bilimoria K, Liu Y, Paruch J, et al. Development and Evaluation of the Universal ACS NSQIP
Surgeons. J Am Coll Surg, 2013;217(5):833-42.
SC
Surgical Risk Calculator: A Decision Aid and Informed Consent Tool for Patients and
M AN U
13. Cohen ME, Bilimoria KR, Ko CY, et al. Development of an American College of Surgeons National Surgery Quality Improvement Program: morbidity and mortality risk calculator for colorectal surgery. J Am Coll Surg, 2009;208(6):1009-16.
14. American Medical Association. CPT 2014 Professional Edition. American Medical Association Press, 2013.
TE D
15. Garrison RN, Cryer HM, Howard DA, Polk HC. Clarification of risk factors for abdominal operations in patients with hepatic cirrhosis. Ann Surg,1984;199(6):648-55. 16. Farnsworth N, Fagan SP, Berger DH, Awad SS. Child-Turcotte-Pugh versus MELD score as a
EP
predictor of outcome after elective and emergent surgery in cirrhotic patients. Am J Surg. 2004;188(5):580-3.
AC C
17. Franzetta M, Raimondo D, Giammanco M, et al. Prognostic factors of cirrhotic patients in extra-hepatic surgery. Minerva Chir,2003;58(4):541-4.
18. Mansour A, Watson W, Shayani V, et al. Abdominal operations in patients with cirrhosis: still a major surgical challenge. Surgery, 1997;122:730–5.
19. Anderson I, Fearon K, Grant I. Laparotomy for abdominal sepsis in the critically ill. Br J Surg, 1996;83(4):535-9.
16
ACCEPTED MANUSCRIPT
20. Ferraris V. Exploratory Laparotomy for Potential Abdominal Sepsis in Patients with Multiple Organ Failure. Arch Surg, 1983;118(10):1130-3. 21. American Hospital Association. Are Medicare Patients Getting Sicker? Trendwatch.
RI PT
December 2012.
22. Tzeng CW, Cooper AB, Vauthey JN, et al. Predictors of morbidity and mortality after hepatectomy in elderly patients: analysis of 7621 NSQIP patients. HPB
SC
(Oxford), 2014;16(5):459-68.
23. Kneuertz PJ, Pitt HA, Bilimoria KY, et al. Risk of morbidity and mortality following hepto-
M AN U
pancreato-biliary surgery. J Gastrointest Surg, 2012;16(9):1727-35.
24. Greenblatt DY, Kelly KJ, Rajamanickam V, et al. Preoperative factors predict perioperative morbidity and mortality after pancreaticoduodenectomy. Ann Surg Oncol, 2011;18(8):2126-35.
25. Englesbe MJ, Patel SP, He K. et al. Sarcopenia and Mortality after Liver Transplantation. J
TE D
Am Coll Surg, 2010;211(2):271-278.
26. Slim K, Panis Y, Alves A, et al. Predicting Postoperative Mortality in Patients Undergoing Colorectal Surgery. World J Surg, 2006;30:100-106.
EP
27. Steiner CA, Bass EB, Talamini MA, et al. Surgical rates and operative mortality for open and LC in Maryland. N Engl J Med 1994;330:403.
AC C
28. Bittner R, Butters M, Ulrich M, et al. Total Gastrectomy. Updated operative mortality and long-term survival with particular reference to patients older than 70 years of age. Ann Surg 1996; 224(1):37-42.
29. Antoniou SA, Antoniou GA, Koch OO, et al. Meta-analysis of laparoscopic versus open repair of perforated peptic ulcer. JSLS 2013;17(1):15-22.
17
ACCEPTED MANUSCRIPT
30. Juo Y, Hyder O, Haider A H, et al. Is Minimally Invasive Colon Resection Better than Traditional Approaches? First Comprehensive National Examination With Propensity Score Matching. JAMA Surg 2014;149(2):177-84.
RI PT
31. Tottrup A, Erichsen R, Svaerke C, et al. Thirty-day mortality after elective and emergency total colectomy in Danish patients with inflammatory bowel disease: a population-based nationwide cohort study. BMJ Open 2012;2(2):e000823.
SC
32. Tapani MJ, Olavi KH. Surgical Management of Toxic Megacolon. Hepatogastroenterology 2014;61(131):638-41.
M AN U
33. Hall JF, Berger D. Outcome of colectomy for Clostridium difficile colitis: a plea for early
AC C
EP
TE D
surgical management. Am J Surg 2008;196(3):384-8.
18
ACCEPTED MANUSCRIPT
Table 1. Patient Characteristics
AC C
EP
TE D
M AN U
SC
RI PT
Patients Hypertension requiring medications 3154 (64.4%) Previous septic shock 3026 (61.7%) Bleeding disorder 1435 (29.3%) Diabetes 1291 (26.3%) History of smoking 1300 (26.5%) History of COPD 984 (20.1%) Concurrent pneumonia 959 (19.6%) Ascites 885 (18.1%) Renal failure 879 (17.9%) Dialysis 784 (16%) Previous SIRS 725 (14.8%) CVA 657 (13.4%) Congestive heart failure 652 (13.3%) Previous major cardiac surgery 642 (13.1%) Steroid 596 (12.2%) Previous sepsis 513 (10.5%) Previous percutaneous coronary intervention 416 (8.5%) MI within 6 months 356 (7.3%) Weight loss (>10% within 6 months) 313 (6.4%) History of alcohol use 301 (6.1%) Peripheral vascular disease 262 (5.3%) Coma 226 (4.6%) Disseminated cancer 214 (4.4%) History of angina 153 (3.1%) Prior chemotherapy 135 (2.8%) Quadriplegia 80 (1.6%) Paraplegia 74 (1.5%) Prior radiation 58 (1.2%) Esophageal varices 44 (0.9%) COPD = Chronic Obstructive Pulmonary Disease. MI = Myocardial Infarction. CVA = Cerebrovascular accident. SIRS = Systemic Inflammatory Response Syndrome
19
ACCEPTED MANUSCRIPT
Table 2. Procedure specific 30-day mortality
51.3% 45.2% 42.2% 28.7% 27.35% 12.5% 51.5% 39.8% 50.0% 54.2%
SC
1570 (32.0%) 1022 (20.9%) 1881 (38.4%) 87 (1.8%) 245 (5.0%) 96 (2.0%) 1521 (31.0%) 3032 (61.9%) 324 (6.6%) 24 (0.5%)
M AN U
Primary procedure Exploratory laparotomy Small bowel Colon Gastric Cholecystectomy Appendectomy Exploratory laparotomy alone One abdominal procedure * Two abdominal procedures* Three or more abdominal procedures*
30-Day Mortality
RI PT
#Patients (% total)
AC C
EP
TE D
*Refers to procedure exclusive of coding for exploratory laparotomy
20
AC C
EP
TE D
M AN U
SC
Table 3. Complications Complication Frequency Failure to wean from ventilator within 48-hours 57.4% Required transfusion 26.6% Pneumonia 16.9% Septic shock 14.2% Re-intubation 11% Acute renal failure 10.4% Sepsis 9.8% Cardiac arrest 7.9% Superficial wound infection 3.5% Wound dehiscence 3.4% Coma 2.3 Deep wound infection 2.1% MI 2% CVA 1.9% Pulmonary emboli 1.2% CVA = Cerebrovascular accident. MI = Myocardial Infarction.
RI PT
ACCEPTED MANUSCRIPT
21
ACCEPTED MANUSCRIPT
Table 4. Outcome measures Average (SD) 25.8 (27.6) days 6.8 (7.9) days 103.7 (68.3) minutes 1.2 (2.5) units 18.8 (21.1) days
AC C
EP
TE D
M AN U
SC
RI PT
Total hospital length of stay Days operation to death Operation time Blood transfusions Days operation to discharge
22
ACCEPTED MANUSCRIPT
Table 5. Significant risk factors predictive of 30-day mortality P-value
Coma
2.63 (1.95-3.55)
P<0.001
Pre-op INR >3.0
2.60 (1.79-3.79)
P<0.001
Esophageal varices
2.44 (1.21-4.91)
P=0.002
Disseminated Cancer
1.98 (1.47-2.66)
P<0.001
Pre-op Creatinine ≥2 mg/dL
1.74 (1.52-1.98)
P<0.001
Pre-op Platelets < 100 (x109/L)
1.72 (1.48-2.00)
Age >65 years
1.70 (1.50-1.93)
Any Sepsis/SIRS
1.57 (1.30-1.90)
Dialysis
1.50 (1.26-1.78)
P<0.001
History of PVD
1.49 (1.14-1.95)
P=0.004
Ascites
1.29 (1.10-1.51)
P=0.002
Pre-op Bilirubin ≥2mg/dL
1.29 (1.10-1.50)
P=0.002
History of COPD
1.23 (1.06-1.44)
P=0.007
Alcohol history
TE D
Wound Infection BMI > 30
SC
RI PT
Odds-ratio (95% Confidence Interval)
P<0.001 P<0.001 P<0.001
M AN U
Risk Factor
0.80 (0.69-0.93)
P=0.003
0.76 (0.69-0.87)
P<0.001
0.73 (0.57-0.96)
P=0.002
AC C
EP
Quadriplegia 0.55 (0.31-0.98) P=0.038 INR = International Normalized Ration. SIRS = Systemic Inflammatory Response Syndrome. PVD = Peripheral Vascular Disease. Na = Serum Sodium. COPD = Chronic Obstructive Pulmonary Disease. BMI = Body Mass Index.
23