Evidence Table

mayberpg

Blunt_small_bowel_perforation2019.pdf

Home >Nursing homework help >Evidence Table

D ow

nloaded from

http://journals.lw w .com

/jtraum a by

7K P ai5G

m 8shhq3O

sB 2blM

9iB ZB

w B f44U

ohC N K x9uK

m 1W

gO q7aufil3w

V K U J9U

0A dTH

V /dn9r7eO

K FZJ56W

f5jq9v4ls715T4uA R G X g4J05A

= on

03/25/2019

Downloadedfromhttp://journals.lww.com/jtraumaby7KPai5Gm8shhq3OsB2blM9iBZBwBf44UohCNKx9uKm1WgOq7aufil3wVKUJ9U0AdTHV/dn9r7eOKFZJ56Wf5jq9v4ls715T4uARGXg4J05A=on03/25/2019

Blunt small bowel perforation (SBP): An Eastern Association for the Surgery of Trauma multicenter update 15 years later

Samir M. Fakhry, MD, Ahmed Allawi, MD, Pamela L. Ferguson, PhD, Christopher P. Michetti, MD, Anna B. Newcomb, PhD, Chang Liu, PhD,

Michelle R. Brownstein, MD, and the EAST small bowel perforation (SBP) Multi-Center Study Group, Reston, Virginia

AAST Continuing Medical Education Article

Accreditation Statement This activity has been planned and implemented in accordance with the Es- sential Areas and Policies of the Accreditation Council for Continuing Medical Education through the joint providership of the American College of Surgeons and the American Association for the Surgery of Trauma. The American College Surgeons is accredited by the ACCME to provide continuing medical education for physicians.

AMA PRA Category 1 Credits™ The American College of Surgeons designates this journal-based CME activity for a maximum of 1 AMA PRA Category 1 Credit™. Physicians should claim only the credit commensurate with the extent of their participation in the activity. Of the AMA PRA Category 1 Credit™ listed above, a maximum of 1 credit meets the requirements for self-assessment.

Credits can only be claimed online

Objectives After reading the featured articles published in the Journal of Trauma and Acute Care Surgery, participants should be able to demonstrate increased understanding of the material specific to the article. Objectives for each article are featured at the beginning of each article and online. Test questions are at the end of the article, with a critique and specific location in the article referencing the question topic.

Claiming Credit To claim credit, please visit the AAST website at http://www.aast.org/ and click on the “e-Learning/MOC” tab. You must read the article, successfully complete the post-test and evaluation. Your CME certificate will be available immediately upon re- ceiving a passing score of 75% or higher on the post-test. Post-tests receiving a score of below 75% will require a retake of the test to receive credit.

Disclosure Information In accordance with the ACCME Accreditation Criteria, the American College of Surgeons, as the accredited provider of this journal activity, must ensure that anyone in a position to control the content of J Trauma Acute Care Surg articles selected for CME credit has disclosed all relevant financial relationships with any commercial interest. Disclosure forms are completed by the editorial staff, associate editors, reviewers, and all authors. The ACCME defines a `commercial interest' as “any entity producing, marketing, re-selling, or distributing health care goods or services consumed by, or used on, patients.” “Relevant” financial relationships are those (in any amount) that may create a conflict of interest and occur within the 12’months preceding and during the time that the individual is engaged in writing the article. All reported conflicts are thoroughly managed in order to ensure any potential bias within the content is eliminated. However, if you’perceive a bias within the article, please report the circumstances on the evaluation form.

Please note we have advised the authors that it is their responsibility to disclose within the article if they are describing the use of a device, product, or drug that is not FDA approved or the off-label use of an approved device, product, or drug or unapproved usage.

Disclosures of Significant Relationships with Relevant Commercial Companies/Organizations by the Editorial Staff Ernest E. Moore, Editor: PI, research support and shared U.S. patents Haemonetics; PI, research support, Instrumentation Laboratory, Inc.; Co-founder, Thrombo Thera- peutics. Associate Editors David Hoyt, Ronald V. Maier and Steven Shackford have nothing to disclose. Editorial staff and Angela Sauaia have nothing to disclose.

Author Disclosures The authors have nothing to disclose.

Reviewer Disclosures The reviewers have nothing to disclose.

Cost For AAST members and Journal of Trauma and Acute Care Surgery subscribers there is no charge to participate in this activity. For those who are not a member or subscriber, the cost for each credit is $25.

System Requirements The system requirements are as follows: Adobe® Reader 7.0 or above installed; Internet Explorer® 7 and above; Firefox® 3.0 and above, Chrome® 8.0 and above, or Safari™ 4.0 and above.

QuestionsQuestions If you have any questions, please contact AAST at 800-789-4006. Paper test and evaluations will not be accepted.

ORIGINAL ARTICLE

642 J Trauma Acute Care Surg

Volume 86, Number 4

http://www.aast.org/

BACKGROUND: Previous work demonstrated diagnostic delays in blunt small bowel perforation (SBP) with increased mortality and inability of scans to reliably exclude the diagnosis. We conducted a follow-up multicenter study to determine if these challenges persist 15 years later.

METHODS: We selected adult cases with blunt injury, International Classification of Diseases, Ninth Revision or current procedural terminology (CPT) indicating small bowel surgery, no other major injury and at least one abdominal computed tomography (CT) within initial 6 hours. Controls had blunt trauma with abdominal CT but not SBP. After institutional review board approval, data from each center were collected and analyzed.

RESULTS: Data from 39 centers (from October 2013 to September 2015) showed 127,919 trauma admissions and 94,743 activations. Twenty- five centers were Level 1. Centers submitted 77 patients (mean age, 39; male, 68%; mean length of stay, 11.3 days) and 131 con- trols (mean age, 44; male, 64.9%; length of stay, 3.6 days). Small bowel perforation cases were 0.06% of admissions and 0.08% of activations. Mean time to surgery was 8.7 hours (median, 3.7 hours). Initial CT showed free air in 31 cases (43%) and none in con- trols. Initial CTwas within normal in three cases (4.2%) and 84 controls (64%). Five cases had a second scan; two showed free air (one had an initial normal scan). One death occurred among the patients (mortality, 1.4%; and time to surgery, 16.9 hours). Regres- sion analysis showed sex, abdominal tenderness, distention, peritonitis, bowelwall thickening, free fluid, and contrast extravasation were significantly associated with SBP.

CONCLUSIONS: Blunt SBP remains relatively uncommon and continues to present a diagnostic challenge. Trauma centers have shortened time to surgery with decreased case mortality. Initial CT scans continue to miss a small number of cases with potentially serious conse- quences. We recommend (1) intraperitoneal abnormalities on CT scan should always evoke high suspicion and (2) strong consid- eration of additional diagnostic/therapeutic intervention by 8 hours after arrival in patients who continue to pose a clinical challenge. (J Trauma Acute Care Surg. 2019;86: 642–650. Copyright © 2019 American Association for the Surgery of Trauma.)

LEVEL OF EVIDENCE: Observational study, level III. KEY WORDS: Blunt small bowel perforation; diagnosis.

This work was presented at the 77th annual meeting of theAmerican Association for the Surgery of Trauma and 4th World Trauma Congress, September 29, 2018, in San Diego, CA.

Small bowel perforation (SBP) is a rare complication of blunt abdominal trauma. Estimates of the frequency of small bowel injury in blunt abdominal trauma range from 5% to 15%,1 with perforation occurring in 1% or less of blunt abdomi- nal traumas.2,3 Reliable diagnosis of SBP remains problematic.4

The difficulty of consistently diagnosing or excluding SBP based on CT scan has been demonstrated. Computed tomography (CT) sensitivity estimates range from 50% to 94%, but CT specificity decreased markedly as sensitivity increased, suggesting different standards in reading the CT scans.1,5–7 Additionally, the highest sensitivity value was calculated based on suggestive injuries; only 50% of the patients met diagnostic criteria for SBP on CT scan.8

Since the advent of CT scan as the primary diagnostic tool for blunt abdominal trauma, delays in the diagnosis of blunt SBP with associated mortality have been documented.2,3,6 Because of the pervasive role of CT scan in the diagnostic evaluation of the modern blunt trauma patient, approximately 15 years ago, we conducted an Eastern Association for the Surgery of Trauma (EAST) multicenter study of blunt hollow viscous injury in the age of CT scan. The study showed that blunt SBP was

uncommon, initial CT scan failed to detect as many as 13% of cases subsequently proven at surgery to have SBP, and that delays in diagnosis of more than 24 hours resulted in in- creased mortality.3,6 Other work suggests mortality increases with delays of more than 8 hours.2

The present study aimed to examine whether increased awareness of diagnostic challenges and/or improved technology since the publication of our 2003 multicenter study has improved diagnostic efficiency and time to surgery, thereby decreasing case mortality. Specifically, we aimed to describe, for patients with blunt abdominal trauma, (1) current diagnostic methods of blunt near-isolated nonduodenal SBP; (2) current mortality and mor- bidity due to blunt near-isolated nonduodenal SBP; (3) the current average time to surgery for these patients; (4) the frequency and use of a second abdominal CT scan, and its association with time to surgery; and (5) incidence of postoperative complications in these patients. Finally, we aimed to determine the sensitivity and specificity of diagnostic findings for blunt near-isolated nonduodenal SBP.

METHODS

Design This was an international multicenter retrospective data

collection study conducted as an approved EAST multi- institutional study in follow-up to our similarly constructed 2003 multicenter study. Hospital-level data were collected on trauma level, continent, presence of a protocol for a second CT in blunt abdominal trauma, number of yearly adult trauma admissions, number of yearly adult trauma activations, number of yearly adult blunt trauma admissions, and approximate num- ber of yearly adult blunt trauma admissions with suspected abdominal injury (i.e., those having CT scan of the abdomen). Patients were identified via medical records for inclusion in the study, and records were screened for the inclusion and

Submitted: September 20, 2018, Revised: December 13, 2018, Accepted: December 14, 2018, Published online: January 10, 2019.

From the Department of Surgery, Reston Hospital Center, Reston, Virginia (S.M.F.); Department of Surgery, Medical University of South Carolina, Charleston, South Carolina (A.A., P.L.F.); Department of Surgery, Inova Fairfax Hospital, Falls Church, Virginia. (C.P.M., A.B.N., C.L.); and Department of Surgery, University of North Carolina, Chapel Hill, NC (M.R.B.).

Address for reprints: Samir M. Fakhry, MD, FACS, Synergy Surgicalists, Department of Surgery, Reston Hospital Center, 1850 Town Center Drive, Suite 309, Reston, VA 20190; email: [email protected].

DOI: 10.1097/TA.0000000000002176

J Trauma Acute Care Surg Volume 86, Number 4 Fakhry et al.

mailto:[email protected]

exclusion criteria (discussed in the next subsection). Data collection from a review of the patients’ charts included pa- tients’ demographics, medical history, injury-related details, physiologic details, diagnostic tests and findings, CT capability (i.e., 32, 64, 128 slice), surgical timing and results, complica- tions, length of stay (LOS), intensive care unit (ICU) LOS, and disposition.

Candidate centers completed an electronic survey to deter- mine eligibility. The coordinating center developed and dissem- inated to the participating centers a data dictionary and detailed data collection protocols. Before beginning data collection, a se- ries of e-mail exchanges and conference calls were conducted to support the participating centers, answer queries, and distribute standardized responses so that all centers were well prepared. De-identified data were submitted from the participating centers to the coordinating center via a secure REDCap database and stored on a password-protected institutional server. Two of the primary investigators (S.M.F. and P.L.F.) reviewed the submitted data files and contacted centers for missing or inconsistent data and related questions. In particular, each submitted case was in- dividually reviewed to ensure adherence to the study protocol and data integrity.

We performed this study after obtaining the approval of and in accordance with our institutional review board guidelines and in compliance with the Health Insurance Portability and Accountability Act. Primary institutional review board ap- proval was obtained at the coordinating center and also from each participating hospital before their data submission.

Subjects We included adult patients (18 years or older) admitted to

participating hospitals between October 1, 2013 and September

30, 2015 with blunt trauma, no duodenal injury (i.e., no Inter- national Classification of Diseases, Ninth Revision, Clinical Modification code of 863.21 or 863.31), and all nonab- dominal Abbreviated Injury Scores (AISs) < 3. The coordi- nating center obtained three groups of patients from each participating trauma center:

1) Potential cases of isolated SBP: patients with an Interna- tional Classification of Diseases, Ninth Revision procedure or CPT code indicating small-bowel surgery with at least one abdominal CT performed within the initial 6 hours from admission;

2) Potential controls: patients with no procedure code indicat- ing small bowel surgery, no AIS greater than 2 including the abdominal region, a trauma activation, and an abdominal CT performed within the initial 6 hours of admission; and

3) Potential SBP diagnosed without CT scan: patients with a procedure code for small-bowel surgery but no abdominal CT done before surgery (the rare case in which a diagnostic technique other than CTwas used).

Once these groups were identified, sites performed further medical record abstraction to confirm that patients in the first and third groups had a surgical report stating the presence of nonduodenal SBP and no other major operative findings requir- ing repair or resection, and confirm patients in the second group did not have abdominal surgery. Each center also selected poten- tial controls from their Trauma Registry by selecting patients from the same time interval as the cases who had no AIS greater than 2 for any injury (abdominal and nonabdominal) and an abdominal CT performed within the initial 6 hours from ad- mission. From this list, each study team attempted to match

TABLE 1. Patient Findings

Variables Levels Control (N = 131)

Case (N = 72) p Unadjusted OR 95% CI p

Age 41 [26–61.25] 38.5 [24–50] 0.240 0.99 0.97-1.00 0.130

Sex Male 85 (64.89%) 49 (68.06%) - - - -

Female 46 (35.11%) 23 (31.94%) 0.760 0.87 0.47-1.60 0.650

Abdominal No 105 (80.15%) 7 (9.72%) - - - -

tenderness Yes 26 (19.85%) 65 (90.28%) <0.001 37.50 15.40-91.32 <0.001

Abdominal No 130 (99.24%) 57 (79.17%) - - - -

distension Yes 1 (0.76%) 15 (20.83%) <0.001 34.21 4.41-265.22 0.001

Peritoneal signs No 130 (99.24%) 50 (69.44%) - - - -

Yes 1 (0.76%) 22 (30.56%) <0.001 57.2 7.51-435.56 <0.001

Seat belt sign No 124 (94.66%) 51 (70.83%) - - - -

Yes 7 (5.34%) 21 (29.17%) <0.001 7.29 2.92-18.22 <0.001

Initial PH Normal 11 (57.89%) 3 (25%) - - - -

Abnormal 8 (42.11%) 9 (75%) 0.140 4.12 0.84-20.28 0.080

Initial base deficit Normal 2 (10.53%) 7 (46.67%) - - - -

Abnormal 17 (89.47%) 8 (53.33%) 0.020 0.13 0.02-0.80 0.030

Initial lactate Normal 21 (48.84%) 9 (31.03%) - - - -

Abnormal 22 (51.16%) 20 (68.97%) 0.150 2.12 0.79-5.70 0.140

Initial WBC Normal 62 (53.45%) 18 (28.57%) - - - -

Abnormal 54 (46.55%) 45 (71.43%) 0.002 2.87 1.49-5.54 0.002

The numbers of missing data records for age, initial PH, base deficit, lactate, and WBC were 1, 172, 169, 131, and 24. CI, Confidence interval; OR, odds ratio; WBC, white blood cell count.

Fakhry et al. J Trauma Acute Care Surg

Volume 86, Number 4

two controls to each case, with age within 5 years and date of admission within 30 days, although not all sites were able to provide that number of controls.

Data Analysis We calculated descriptive statistics on hospital level data.

Descriptive statistics were used to describe current methods trauma centers are using to diagnose near-isolated nonduodenal SBP in patients with blunt abdominal trauma in conjunction with abdominal CT, focused assessment with sonography in trauma, abdominal ultrasound, x-ray, and/or diagnostic perito- neal lavage. Percentages of positive findings for each diagnostic modality, as well as physical examination, laboratory values, and vital signs, were calculated for patients with SBP and those without SBP.

To describe current mortality and morbidity due to blunt near-isolated nonduodenal SBP, we calculated overall mortality rate of patients with SBP, mortality rates by time to surgery, and morbidity rate of types of complications. To describe the current average time to surgery for patients with blunt near- isolated nonduodenal SBP, we calculated overall time to surgery, as well as percentages of patients with SBP going to surgery in less than 8 hours, less than 16 hours, less than 24 hours, and 24 hours or more. To describe the frequency and use of a second

abdominal CT scan, and its association with time to surgery and incidence of postoperative complications in patients with blunt near-isolated nonduodenal SBP who did not go to sur- gery emergently (within 4 hours) after an initial CT, we calcu- lated the percentage of patients with SBP who had a second

TABLE 3. Initial CT Scan Findings

Findings Levels Control Case p Unadjusted OR 95% CI p

Free air* No 131(0%) 41 (56.94%) - - - -

Yes 0 (0%) 31(43.06%) <0.001 - - -

Free fluid No 120 (91.6%) 12 (16.67%) - - - -

Yes 11 (8.4%) 60 (83.33%) <0.001 54.55 22.74-130.83 <0.001

Bowel wall thickening

No 130 (99.24%) 45 (62.5%) - - - -

Yes 1 (0.76%) 27 (37.5%) <0.001 78 10.31-590.03 <0.001

Mesenteric stranding

No 130 (99.24%) 55 (76.39%) - - - -

Yes 1 (0.76%) 17 (23.61%) <0.001 40.18 5.22-309.41 <0.001

Contrast extravasation

No 130 (99.24%) 64 (88.89%) - - - -

Yes 1 (0.76%) 8 (11.11%) 0.001 16.25 1.99-132.73 0.009

Solid organ injury

No 119 (90.84%) 67 (93.06%) - - - -

Yes 12 (9.16%) 5 (6.94%) 0.79 0.74 0.25-2.19 0.590

Retroperitoneal blood

No 126 (96.18%) 66 (91.67%) - - - -

Yes 5 (3.82%) 6 (8.33%) 0.2 2.29 0.67-7.79 0.180

Chance fracture

No 129 (98.47%) 70 (97.22%) - - - -

Yes 2 (1.53%) 2 (2.78%) 0.62 1.84 0.25-13.37 0.550

* Due to complete separation, unadjusted OR and 95% CI were not calculated for free air.

TABLE 2. Frequencies of Diagnostic Testing Used

Test All

(N = 203) Control (n = 131)

Case (n = 72)

CT scan 203 (100%) 131 (100%) 72 (100%)

Chest x-ray 159 (78.33%) 105 (80.15%) 54 (75%)

FAST 81 (39.9%) 52 (39.69%) 29 (40.28%)

Abdominal x-ray 14 (6.9%) 6 (4.58%) 8 (11.11%)

DPL 0 (0%) 0 (0%) 0 (0%)

DPL, Diagnostic peritoneal lavage; FAST, focused assessment with sonography for trauma.

TABLE 4. Sensitivities and Specificities of the Diagnostic Findings

Sensitivity Specificity PPV NPV Accuracy

Physical examination at admission

Abdominal tenderness

90.28% 80.15% 71.43% 93.75% 83.74%

Abdominal distention

20.83% 99.24% 93.75% 69.52% 71.43%

Peritoneal signs 30.56% 99.24% 95.65% 72.22% 74.88%

Seat belt sign/bruising

29.17% 94.66% 75.00% 70.86% 71.43%

Other 18.06% 71.76% 26.00% 61.44% 52.71%

Initial laboratory results (abnormal)

Arterial pH 75.00% 57.89% 52.94% 78.57% 64.52%

Base deficit 53.33% 10.53% 32.00% 22.22% 29.41%

Lactate, mmol/L 68.97% 48.84% 47.62% 70.00% 56.94%

WBC count 71.43% 53.45% 45.45% 77.50% 59.78%

CT findings (initial)

Free fluid 83.33% 91.6% 84.51% 90.91% 88.67%

Free air 43.06% 100% 100% 76.16% 79.80%

Bowel wall thickening

37.50% 99.24% 96.43% 74.29% 77.34%

Mesenteric stranding

23.61% 99.24% 94.44% 70.27% 72.41%

Contrast extravasation

11.11% 99.24% 88.89% 67.01% 67.98%

Solid organ injury 6.94% 90.84% 29.41% 63.98% 61.08%

Retroperitoneal blood

8.33% 96.18% 54.55% 65.62% 65.02%

Chance fracture 2.78% 98.47% 50.00% 64.82% 64.53%

J Trauma Acute Care Surg Volume 86, Number 4 Fakhry et al.

abdominal CT scan before surgery. We calculated the sensi- tivity, specificity, positive predicative value (PPV), negative predicative value (NPV), and accuracy for each individual finding from physical examination, initial CT scan, and labo- ratory test results. Unadjusted odds ratios (ORs) were calcu- lated from univariate logistic regression models along with 95% confidence intervals for demographic factors and diagnos- tic findings. After stepwise regression (forward and backward) including covariates that had p < 0.1 from univariate analysis, a multivariate logistic regression model was used to obtain ad- justed ORs. We calculated descriptive statistics on hospital level data and rate of false-negative CT scans.

RESULTS

Centers and Cases Datawere available from 39 centers** with 127,919 trauma

admissions and 94,743 trauma activations from October 2013 to September 2015. Twenty-nine centers were Level 1 trauma cen- ters, 7 were Level 2, 1 was Level 3, and 2 were international. Thirty-three centers (85%) were able to identify and submit 77 SBP potential cases (Groups 1 and 3). The SBP patients were 0.06% of admissions and 0.08% of activations. There were 77 patients enrolled (mean age, 39 years; male, 68%; mean hos- pital LOS, 11.3 days) and 131 non-SBP controls (mean age, 44 years; male, 64.9%; LOS, 3.6 days). Of the 77 patients with near-isolated nonduodenal SBP, 5 underwent surgery without a CT scan and were not included in subsequent analyses (did not meet inclusion criteria). The subsequent analysis considered the 72 enrolled patients with proven SBP at surgery and a qualifying CT scan as cases. The median number of cases each center con- tributed was 2 (interquartile range [IQR], 1–3). Fifteen centers had only one case, five centers had two cases, eight centers had three cases, two centers had four cases, and three centers had five

cases. Overall, the median injury severity score (ISS) was 8, IQR [5–10]; median ISS for cases was 10, IQR [9–14] and median ISS for controls was 5, IQR [4–9]. The most common mechanism of injury by far was motor vehicle crash (65% of cases and 50% of controls) followed by falls (9.7% of cases and 15% of controls) and motorcycle crashes (5.6% for cases and 10.7% for controls).

Diagnostic Testing Patient findings are shown in Table 1. On univariate analy-

sis, cases (patients with proven SBP and qualifying preoperative CT scan) were significantly more likely than controls (patients without SBP) to have abdominal tenderness, abdominal disten- tion, a seat belt sign, peritoneal signs, abnormal base deficit, and elevated white blood cell count. Initial pH and lactate levels were not significantly associated with the finding of SBP. Very limited conclusions should be drawn involving the laboratory results because of significant missing data.

The frequency of diagnostic testing used among those pa- tients with an initial CT scan is shown in Table 2. All controls and 72 of the 77 cases had an initial CT scan. Initial CT scan showed free air in 31 cases (43%) and none of the controls (Table 3). Free fluid (in most cases without solid organ injury) was present in 60 cases (83.3%) but only 11 controls (8.4%). Bowel wall thickening, mesenteric stranding, and contrast ex- travasation were almost never seen in controls and occurred at varying and statistically significant rates (37.5%, 23.61%, and 11.11%, respectively) in cases. The unadjusted ORs for free fluid, bowel wall thickening, and mesenteric stranding were 54, 78, and 40, respectively.

Initial CT scan was reported as within normal limits in three case patients (4.2%) and 84 controls (64%). The three nor- mal CT scans in case patients were performed in machines with 64 or more detectors. Two of the three patients whose initial CT scan was normalwent on to surgery without further imaging, while the third had a second CT scan that showed free air prompting ex- ploratory laparotomy. An additional four case patients had a sec- ond CT scan (for a total of five case patients having a second CT scan). One of the four follow-up CT scans showed free air; thus, two of the five second CT scans showed free air.

The sensitivity and specificity of diagnostic findings in initial CT scans are shown in Table 4. Free air on CT scan was highly specific (100% PPV) but only somewhat sensitive (43% NPV) for SBP. The sensitivity and specificity values for

TABLE 5. Regression Analysis

OR 95% CI p

Multivariate Logistic Regression Model

Age 0.98 0.93-1.03 0.427

Sex (female vs. male) 0.1 0.01-0.85 0.035

Abdominal tenderness 46.1 5.33-398.88 0.001

Abdominal distention 443.92 1.73-114068.66 0.031

Peritoneal signs 268.81 6.47-11159.85 0.003

Seatbelt signs 2.76 0.37-20.59 0.323

Free fluid 93.39 11.8-738.93 <0.001

Bowel wall thickening 583.36 10.85-31372.37 0.002

Mesenteric stranding 24.49 0.02-33496.01 0.385

Contrast extravasation 87.12 1.06-7172.57 0.047

Stepwise Regression Model

Sex (female vs. male) 0.14 0.02-0.89 0.037

Abdominal tenderness 45.59 5.99-347.08 <0.001

Abdominal distention 434.2 3.91-48219.57 0.011

Peritoneal signs 174.91 5.91-5176.21 0.003

Free fluid 93.3 13.31-654.2 <0.001

Bowel wall thickening 751.46 13.53-41737.43 0.001

Contrast extravasation 29.64 0.82-1070.24 0.064

TABLE 6. Complications

Complication All Patients Controls Cases p

Wound infection 3 (1.44%) 0 (0%) 3 (3.9%) 0.049

Abscess 4 (1.92%) 0 (0%) 4 (5.19%) 0.018

Sepsis 4 (1.92%) 0 (0%) 4 (5.19%) 0.018

Pneumonia 6 (2.88%) 2 (1.53%) 4 (5.19%) 0.197

ARDS 4 (1.92%) 1 (0.76%) 3 (3.9%) 0.144

AKI 4 (1.92%) 0 (0%) 4 (5.19%) 0.018

VTE 2 (0.96%) 0 (0%) 2 (2.6%) 0.136

Unplanned return to surgery 4 (1.92%) 0 (0%) 4 (5.19%) 0.018

Other 18 (8.65%) 3 (2.29%) 15 (19.48%) <0.001

AKI, acute kidney injury; ARDS, acute respiratory distress syndrome; VTE, venous thromboembolism.

Fakhry et al. J Trauma Acute Care Surg

Volume 86, Number 4

free fluid on CT scan were both relatively high, resulting in the highest overall accuracy (88.67%) for the diagnosis of SBP. Other CT abnormalities such as bowel wall thickening, mesen- teric stranding, and contrast extravasation had high specificity but low sensitivity rates with resultant lower accuracy rates.

Multivariate logistic regression analysis and stepwise re- gression (for variable selection) was performed incorporating demographic factors, physical examination results, and CT scan findings (Table 5). Stepwise regression model dropped both seatbelt sign and mesenteric stranding from the model. Both models showed that female sex, abdominal tenderness, abdominal distention, peritoneal signs, free fluid, and bowel wall thickening were significantly associated with SBP. Contrast extravasation reached significance in the first model but was no longer signifi- cant in the second, more parsimonious model.

Time to Surgery Mean time to surgery was 8.43 hours (median, 3.74 hours;

IQR, 2.00–10.53). Most cases (52 patients, 78%) went to sur- gery within 8 hours. Nine patients (12.5%) went to surgery be- tween 8 and 16 hours, four patients (5.6%) went between 16 and 24 hours, and seven patients (9.7%) went after the first 24 hours.

The five patients who had a second CT scan went to sur- gery 22, 13.9, 9.7, 25.5, and 4.5 hours after admission. All five survived and were discharged home. Of the three patients whose initial CT scan was negative, two went to surgery 50.3 and 10.3 hours after admission, while the third had a second CT scan then went to surgery 22 hours after admission (shown as one of the five cases previously mentioned). All three of these patients survived and went home.

Complications and Mortality Complications were uncommon in the control patients. A

number of complications occurred significantly more often in case patients including wound infection, abscess, sepsis, acute kidney injury, unplanned return to surgery, and others (Table 6). Pneumonia, acute respiratory distress syndrome, and venous thromboembolism occurred more frequently in cases than in controls, but the rates did not attain statistical significance likely owing to the relatively small numbers encountered.

One death occurred among the case patients for an overall mortality rate of 1.4%. The patient who died was a 71-year-old male with an admission Glasgow Coma Scale score of 15 and an ISS of 8. His initial temperature was normal, his heart rate was 66 beats per minute, and his white blood cell count was 4,000. This patient's initial CT scan was performed 93 minutes after admission and showed free fluid. He went to surgery 16.9 hours after admission.

DISCUSSION

Conducted 15 years after our initial EAST multicenter study of hollow viscous injury,3,6 the present study confirmed our previous findings that SBP is rare (0.06% of admissions and 0.08% of activations in this study), CT scans fail to detect any abnormality in a small but significant number of patients with proven SBP (4.2%), and delays to surgery are associated with mortality in some patients. There seems to have been

progress, however, in several areas: a decrease in the rate of false-negative CT scans from 13% to 4.2%; a decrease in case mortality from 6% to 1.3%, and shortened time to surgery. De- spite these improvements, a patient who went to surgery 16.9 hours after admission died of complications after what is essentially a simple injury to repair.9

The ability of physical examination to reliably diagnose SBP continues to be limited. In this study, the absence of perito- neal signs did not exclude SBP, as 69.4% of patients with proven SBP were reported as not having peritonitis on presentation. The limitations of physical examination findings were highlighted in our 2003 study,6 and the recent literature supports this inability to rely solely or primarily on physical examination.10–12 Caution should be exercised in interpreting these data, however, as phys- ical findings alone may be unreliable for the accurate diagnosis of SBP, but abdominal examination remains a key element in de- cision making for operative intervention in nearly all studies. This is supported by the appearance of abdominal examination ele- ments in most recommendations and regression-based models for prediction of this injury, including the one developed in the present study.6,11,12

The use of CT scan in hemodynamically stable blunt trauma patients at risk of intra-abdominal injury remains ubiqui- tous, and as such, CT scan findings receive special consideration in determining whether patients have important intra-abdominal injuries. The presence of free air on CT scan is perhaps the most widely accepted indication for exploratory laparotomy in a he- modynamically stable patient. It should be noted, however, that although it was present in none of the patients without SBP, it was present in only 43% of patients with proven SBP. Thus, free air on CT scan is highly specific (100% PPV) but only some- what sensitive (76% NPV) for SBP; stated differently, it is a suf- ficient but not necessary condition for the presence of SBP, and its absence does not exclude the diagnosis. Free fluid (especially without solid organ injury) was significantly more likely to be present in patients with SBP than in those without. The sensitiv- ity and specificity values for free fluid on CT scan were both rel- atively high, resulting in the highest overall accuracy (88.67%) for the diagnosis of SBP. Other CT abnormalities such as bowel wall thickening, mesenteric stranding, and contrast extravasation occurred at varying rates in cases but were almost never seen in controls. As such, the presence of these findings on CT scan should raise suspicion of the presence of SBP. Several recent studies also advocate increased scrutiny of patients with CT scan abnormalities such as bowel wall thickening and mesen- teric stranding because of their frequent association with bowel injury.12–14

Based on the results of this study and several recent pub- lications, the limitations of CT scan in the diagnosis of patients with SBP should be re-emphasized. Despite improvements in technical quality of CT scanners and better identification of CT findings that raise the suspicion of SBP,10,14 reports of false-negative CT scan results continue to appear. In a single- center study, Becker et al. reported a 28% rate of negative CT scans in patients ultimately proven to have SBP.11 In another single-center study, McNutt et al. reported that as many as 25% of the patients they selected for nonoperative management failed that regimen and required surgery, suggesting that their initial CT scans were insufficiently precise for the diagnosis.12

J Trauma Acute Care Surg Volume 86, Number 4 Fakhry et al.

As a result, reliance on CT scan to exclude this uncommon diag- nosis must be tempered with the knowledge drawn from numer- ous studies that false-negative rates of CT scan for SBP continue to be significant even with advanced scanner technology and ex- pert interpretation with rates from 4.2% to 28%. Assuming a 5% miss rate on initial CT scan (a low estimate) and a prevalence of SBP of 0.5% of activations (a high estimate), a trauma center that has 1,500 activations a year will see 22.5 patients with SBP every 3 years and can expect to miss the diagnosis on initial CT scan in approximately one patient. It is worth noting that in the present study, there were five centers (15%) with no cases over the 2-year period of the study, confirming the low rates at which this injury is encountered.

Modalities other than CT scan have less established roles for the diagnosis of blunt SBP. Ultrasonography is useful when patients have intraperitoneal free air or fluid, but is not sensitive in the absence of those markers.15 Laparoscopy and peritoneal lavage are the most sensitive methods, with most sensitivity measured at 100%, but laparoscopy especially may not be feasi- ble in all patients and under all circumstances.1,5,16,17 Although SBP is clearly a time-sensitive injury, these data suggest that short delays (likely <8 hours) to use the more sensitive diagnos- tic method of laparoscopy could be beneficial, rather than rely- ing solely on CT scan.

Since increased delay in surgical repair is associated with increased morbidity and mortality, improving sensitivity of existing examination methods is critical. Mortality from hollow viscous in- jury has been shown to increase from 13.0% with surgery within 8 hours to 30.8% when surgery is delayed for more than 24 hours.6

In addition, delay in surgery results in increased likelihood of wound infection, wound dehiscence, intra-abdominal abscess, acute respiratory distress syndrome, and sepsis.6 Delay in sur- gery past 24 hours has also been associated with increased hos- pital stay.5 The mortality in the present study occurred despite surgery occurring at 16.9 hours from admission. This is consis- tent with data that suggest that surgery in less than 24 hours is necessary in these patients, with surgery in less than 8 hours preferred.2,5 The rarity of this type of injury, combined with the low sensitivity of many diagnostic tests and findings, makes this a difficult diagnosis and, as a result, surgery may be significantly delayed for as long as 50 or more hours from admission as noted in the present study and others as well. Faria et al. noted that all patients with SBP in their study who ultimately died (n = 5) had surgery performed after the first 24 hours (median, 48 hours) compared to 4 hours for the patients who survived (p < 0.001).10 It is clear that delays in the diagnosis of SBP can result in serious complications and death, making timely diagnosis critical in the treatment of these patients. It re- mains to be determined why some patients suffer serious com- plications and even death with surgical intervention occurring within the 24-hour window, while others survive despite pro- longed delays. Additional research into individual variables such as phenotypic expression of genomic characteristics in response to peritoneal contamination and sepsis is needed to explain the observed variations in pathophysiology and outcomes.

This work strongly suggests that progress has been made in the 15 years since our 2003 study. It is not clear if this is because of better CT scan technology and/or interpretation or heightened clinical acumen with a shortened time to surgery or

a combination of all of these. Other less obvious factors may also have contributed to this apparent improvement in outcomes. It is also unclear whether we will be able to completely eliminate mortality from this relatively simple injury that can be readily repaired surgically. Individual patient factors may render a small number of patients exquisitely sensitive to the inflammatory re- sponse triggered by peritoneal contamination.

Our research has a number of limitations. This was an un- funded study that relied on the resources available at each of the participating centers. Despite our extensive efforts to standardize data retrieval and collection, there may have been deviations from the protocol that we could not identify and remedy. The controls in this study were not tightly matched to the cases be- cause of limitations of numbers and resources, and this could bias the results of the study. Due to the retrospective nature of the re- search, differing approaches and interpretations of findings (e.g., reading of CT scans) likely contribute to significant variation in results. Because SBP is an uncommon condition, extremely large numbers of cases would be needed to ensure that our conclusions are sound. This is difficult to accomplish without funding, and it is unlikely that significant funding could be secured for this pur- pose. This nonrandom sample of trauma centers is not entirely representative of the universe of facilities where injured patients are treated, and this may introduce occult biases. Subtle differ- ences in methodology prevent exact comparison to the earlier study. Finally, the ultimate decision to operate upon an injured pa- tient remains a subjective one, and the impact of differing surgical decision making and the subsequent effects on diagnostic choices and patient outcomes cannot be accounted for by this research.

CONCLUSION

Blunt small bowel perforation continues to present a clin- ically challenging triad: a relatively rare injury with an often in- sidious presentation (and therefore little accumulated clinical experience); a small but significant incidence of missed diag- nosis by CT scan, the most relied on diagnostic modality; and serious consequences including morbidity and increased mor- tality when the diagnosis is delayed. This study demonstrates that blunt SBP remains relatively uncommon and continues to present a diagnostic challenge 15 years after our initial multi- center study. Trauma centers seem to have shortened time to surgical intervention with a gratifying associated decrease in case mortality. Initial CT scans continue to miss a small but sig- nificant number of cases with potentially serious consequences making heightened awareness of this injury and continued clin- ical vigilance paramount. Until more definitive solutions are available, we recommend that (1) intraperitoneal abnormalities on CT scan should always evoke high suspicion and (2) strong consideration of additional diagnostic or therapeutic interven- tion by 8 hours after arrival in patients who continue to pose a clinical challenge.

AUTHORSHIP

SMF developed the study concept. SMF, PLF, and AA participated in the study design; all authors performed data collection. SMF, AA, PLF, ABN, and CL collated and analyzed the data. All authors participated in data in- terpretation. SMF and PLF wrote the manuscript; SMF, AA, PLF, CPM, ABN, and MRB performed critical review and revisions.

Fakhry et al. J Trauma Acute Care Surg

Volume 86, Number 4

ACKNOWLEDGMENT

Study data were collected and managed using REDCap (Research Elec- tronic Data Capture) electronic data capture tools, hosted at the Medical University of South Carolina through grant UL1 TR001450.

DISCLOSURE

The authors declare no conflicts of interest. No funding was provided for this study.

REFERENCES 1. Mathonnet M, Peyrou P, Gainant A, Bouvier S, Cubertafond P. Role of lap-

aroscopy in blunt perforations of the small bowel. Surg Endosc. 2003;17(4): 641–645.

2. Fakhry SM, Brownstein M, Watts DD, Baker CC, Oller D. Relatively short diagnostic delays (<8 hours) produce morbidity and mortality in blunt small bowel injury: an analysis of time to operative intervention in 198 patients from a multicenter experience. J Trauma. 2000;48(3):408–414; discussion 414–405.

3. Watts DD, Fakhry SM. EAST Multi-Institutional Hollow Viscus Injury Re- search Group. Incidence of hollow viscus injury in blunt trauma: an analysis from 275,557 trauma admissions from the EAST multi-institutional trial. J Trauma. 2003;54(2):289–294.

4. Mukhopadhyay M. Intestinal injury from blunt abdominal trauma: a study of 47 cases. Oman Med J. 2009;24(4):256–259.

5. Fang JF, Chen RJ, Lin BC, Hsu YB, Kao JL, Kao YC, Chen MF. Small bowel perforation: is urgent surgery necessary? J Trauma. 1999;47(3):515–520.

6. Fakhry SM, Watts DD, Luchette FA. EAST Multi-Institutional Hollow Vis- cus Injury Research Group. Current diagnostic approaches lack sensitivity in the diagnosis of perforated blunt small bowel injury: analysis from 275,557 trauma admissions from the EAST multi-institutional HVI trial. J Trauma. 2003;54(2):295–306.

7. Zissin R, Osadchy A, Gayer G. Abdominal CT findings in small bowel per- foration. Br J Radiol. 2009;82(974):162–171.

8. Sherck J, Shatney C, Sensaki K, Selivanov V. The accuracy of computed to- mography in the diagnosis of blunt small-bowel perforation. Am J Surg. 1994;168(6):670–675.

9. Coleman JJ, Zarzaur BL. Surgical management of abdominal trauma: hollow viscus injury. Surg Clin North Am. 2017;97:1107–1017.

10. Faria GR, Almeida AB, Moreira H, Barbosa E, Correia-da-Silva P, Costa-Maia J. Prognostic factors for traumatic bowel injuries: killing time. World J Surg. 2012;36:807–812.

11. Bekker W, Kong VY, Laing GL, Bruce JL, Manchev V, Clarke DL. The spec- trum and outcome of blunt trauma related enteric hollow visceral injury. Ann R Coll Surg Engl. 2018;100:290–294.

12. McNutt MK, Chinapuvvula NR, Beckmann NM, Camp EA, Pommerening MJ, Laney RW, West OC, Gill BS, Kozar RA, Cotton BA, et al. Early surgical in- tervention for blunt bowel injury: the Bowel Injury Prediction Score (BIPS). J Trauma Acute Care Surg. 2015;78(1):105–111.

13. Firetto MC, Sala F, Petrini M, Lemos AA, Canini T, Magnone S, Fornoni G, Cortinovis I, Sironi S, Biondetti PR. Blunt bowel and mesenteric trauma: role of clinical signs along with CT findings in patients' management. Emerg Radiol. 2018;25:461–467.

14. LeBedis CA, Anderson SW, Soto JA. CT Imaging of blunt traumatic bowel and mesenteric injuries. Radiol Clin North Am. 2012;50:123–136.

15. Moriwaki Y, Sugiyama M, Toyoda H, Kosuge T, Arata S, Iwashita M, Tahara Y, Suzuki N. Ultrasonography for the diagnosis of intraperitoneal free air in chest-abdominal-pelvic blunt trauma and critical acute abdominal pain. Arch Surg. 2009;144(2):137–141.

16. Ceelen W, Hesse U, De Hemptinne B. Small bowel perforation following blunt abdominal trauma. Acta Chir Belg. 1995;95(Suppl 4):187–189.

17. Nicolau AE, Tuţuianu R, Veste V, Merlan V, Negru M, Florescu R, Paizis E. Small bowel perforation caused by compound pelvic fracture found in diag- nostic laparoscopy. Chirurgia (Bucur). 2006;101(4):423–428.

J Trauma Acute Care Surg Volume 86, Number 4 Fakhry et al.

APPENDIX Complete Author List, Alphabetical Order by Hospital Name

John Berne, MD, Dalier R. Mederos, MD, Broward Health Medical Center Dirk Stengel, MD, PhD, MSc, Moritz Weigeldt, MD, Center for Clinical Research, BG Klinikum Unfallkrankenhaus Berlin Forrest O'Dell Moore III, MD, Jeanette G. Ward, MS-CR, Chandler Regional Medical Center Charles E Frank, MD, Christopher J. Dente, MD, Emory University at Grady Memorial Hospital Jeffrey LWild, MD, Katelyn A. Young, BS, Geisinger Medical Center Javier Martin Perez, MD, Hackensack University Medical Center Dennis Y Kim, MD, Erin Howell, MD, Harbor-UCLA Medical Center Karyn L Butler MD, D'Andrea Joseph MD, Hartford Hospital Ashley P Marek, MD, Rachel M Nygaard, PhD, Hennepin County Medical Center Christopher P Michetti, MD, Erik Teicher, MD, Anna B Newcomb, PhD, MSW, Chang Liu, PhD, Inova Fairfax Medical Center Therese M Duane, MD, MBA, Celeste Caliman, RN, John Peter Smith Health Network Lisa M Kodadek, MD, Elliott R Haut, MD, PhD, Johns Hopkins Bayview Lisa M Kodadek, MD, Joseph V Sakran, MD, Johns Hopkins Hospital Marcela C Ramirez, MD, Arianne M Gallaty, MD, Kendall Regional Medical Center Rovinder Sandhu, MD, Christine Campbell, MS, BSN, Lehigh Valley Health Network Benjamin Moran, MD, Thomas A Santora, MD, Lewis Katz School of Medicine at Temple University Richard D Catalano, MD, James Andersen, MD, Loma Linda University Medical Center Fred Luchette, MD, Michael J. Anstadt, MD, Loyola University Medical Center Christopher M Dodgion, MD, MSPH, MBA, Robert K. McMillan MD, Medical College of Wisconsin Samar M Fakhry, MD, Pam L Ferguson, PhD, Ahmed Allawi, MD, Medical University of South Carolina Thomas Schroeppel, MD, MS, Memorial Hospital Colorado Springs Daniel Cullinane, MD, Ministry St Joseph's TC/Marshfield Clinic Elizabeth Acquista, MD, William F. Powers IV, MD, New Hanover Regional Medical Center Carrie Garcia, RN-BSN, Melvin E. Stone Jr., MD, NYC Health and Hospitals/Jacobi Collin E M Brathwaite, MD, Patrizio Petrone, MD, MPH, MSHSA, NYU Winthrop Hospital Stephanie Montgomery, MD, Laura Carrihill Lamb, MD, Saint Francis Hospital and Medical Center Justin L Regner, MD, Lorie Fares, Scott & White Medical Center Brian C Thurston, MD, Amy V. Hamrick, MSN, Spartanburg Medical Center Ariel Santos, MD, MPH, Texas Tech University Health Sciences Center Maher M Matar, MD, MHA, Alexandre Tran, MD, The Ottawa Hospital Rebecca Schroll, MD, Nicholas J Buffin, MD, Tulane University School of Medicine Julie A Dunn, MD, MS, Katherine Corey, BSN, UCHealth North Medical Center of the Rockies Andrew R Doben, MD, Ronald I Gross, MD, UMMS-Baystate Michelle R Brownstein, MD, Megan T. Quintana, MD, University of North Carolina at Chapel Hill Jeffry Nahmias, MD, MHPE, Cesar Figueroa, MD, University of California, Irvine Jason L Sperry, MD, MPH, Vincent PAnto, University of Pittsburgh Medical Center James M Haan, MD, Kelly L Lightwine, MPH, Via Christi Hospital St. Francis Pascal O Udekwu, MD, WakeMed Health & Hospitals Uzer Khan, MD, Nicole Cornell, MS, West Virginia University.

Fakhry et al. J Trauma Acute Care Surg

Volume 86, Number 4