The Accuracy of POCUS in Small Bowel Obstruction

SBO

Accuracy of abdominal ultrasound for the diagnosis of small bowel obstruction in the emergency department

World Journal of Emergency Medicine 2018 - Pubmed Link


Other Articles Referenced in this Podcast

Systematic Review and Meta-analysis on Diagnostics in SBO 2013

Systematic Review and Meta-analysis of Ultrasound in SBO 2018


Take Home Points

1. This study is small and has too many biases to take the accuracies presented at face value. We need larger, prospective studies.

2. In combination with prior studies, this study supports the concept that point of care ultrasound can accurately diagnose SBO. In this case, it showed high specificity and sensitivity in a high pre-test probability population.


Background

Small bowel obstruction (SBO) is probably one of the more underutilized indications for point of care ultrasound. Part of this is probably because not many people realize you can do it and also because it has not been well studied. However, the evidence that exists seems to indicate that this is way better than abdominal xrays (duh) and even approaches the accuracy of CT. Clearly, ultrasound has a lot of advantages over CT in this situation. You can perform it quickly at the bedside, having your answer hours before the patient finally gets their CT. This can theoretically help expedite management. Maybe...just maybe, we could avoid doing the CT. Right now we would need a lot more evidence to back this up. In the absence of robust data on accuracy, this article takes another look to see how accurate US is compared to CT specifically in the emergency department with with a variety of POCUS users.


Question

How accurate is POCUS for diagnosing small bowel obstruction?


Population

Tertiary care academic medical center

Emergency department patients


Inclusion:

  • Had a point of care ultrasound performed for SBO + had a CT performed

Exclusion:

  • Transferred with imaging confirming diagnosis of SBO


Design

Retrospective, single center cohort study

Ultrasound database was queried to find abdominal studies looking for small bowel obstruction

Those patient encounters were reviewed to see if CT was performed during that visit

Considered positive if radiology read CT as positive for SBO (including “early” or “partial”) or if SBO was discharge diagnosis.

Accuracy of POCUS for SBO was calculated


Who did the ultrasounds?

Residents, physician assistants, and ultrasound fellows

All were taught how to perform a protocol to assess for small bowel obstruction, but no specific teaching for this study


The Scan

Curvilinear transducer

Curvilinear


Study Scanning Protocol (to be performed in suspected SBO)

1/ All four quadrants scanned with curvilinear at depth 12-18 cm

2/ Save video clips of peristalsis and still images with measurements of bowel in all quadrants

3/ Criteria to diagnose SBO: 1) Dilation >2.5 cm fluid filled loops of bowel AND 2) Abnormal back and forth peristalsis.


5minsono

5 Minute Sono - Small Bowel Obstruction


thePOCUSAtlas

The POCUS Atlas - Bowel

The Evidence Atlas - Bowel


Results

N = 47

  • 64 patients had POCUS for SBO

  • 9 patients had no CT → excluded

  • 8 patients had indeterminant US findings → excluded


The Patients

  • mean age 58.8 years old

  • 34.4% male

  • 50% prior SBO

  • 85.9% history of prior surgery

  • 62.5% had active malignancy

  • 10.9% had surgery within 2 weeks of ED presentation

  • 50% had small bowel obstruction diagnosed

  • 88% were treated conservatively, 12% had surgery


Primary Outcome Accuracy of Point of care US for SBO in the ED

Sensitivity of 93.8% (79.2-99.2%)

Specificity of 95.2% (76.2-99.9%)

+LR 14.1 (2.11-93.6)

-LR 0.07 (0.02-0.26)


Additional Findings

Who performed these scans?

  • Emergency medicine residents 58%

  • Physician assistants 22%

  • Ultrasound fellows/faculty 20%

Note: they did not do subgroup analysis given low n


Most common alternate diagnoses found on CT

  • 22.7% increased tumor burden

  • 18.2% no explanation

  • 13.6% enteritis/colitis


What happened with the "POCUS Indeterminant" patients?

  • 1 had SBO, 7 did not


Three patients that did not have a CT, had a discharge diagnosis of SBO (a clinical diagnosis!? Incredible!)

If using endpoint of CT + discharge diagnosis for comparison (n = 56), accuracy improves

Sensitivity 94.3%

Specificity 95.2%

+LR 19.8

-LR 0.06


Limitations

Small population, retrospective, single center. There are a lot of biases that could potentially affect this study. One problem was that the participants were identified from only those who received ultrasounds, not all those who had SBO. What makes the provider perform the ultrasound? Because they are good at it? Because it is so clinically obvious that they think they can expedite management? Factors such as patient acuity, patient body habitus, volume of the ED that day would affect who received an ultrasound and that can alter the results significantly. Also this department had a lot of cancer patients. The incidence of the disease can also affect the accuracy of these test results so this is important to external validity. Obviously, a prospective study with a clearly defined protocol would be a better option.

Variety of skill levels in the performers of the ultrasound. 13% of studies were indeterminant. This could be a strength because it might allow this data to apply more broadly to different operators.

Excluded indeterminant exams. These were scans that usually had only one of the two required sonographic features of SBO. Excluding exams that are inconclusive can be a way to make your accuracy look better than it actually is in real life. Fortunately, they let us know what happened with these patients. One actually had an SBO, 7 did not. Although this is too small a population to draw conclusions from, in this study, when it was equivocal, it was usually not a true obstruction.


Discussion

Authors suggest that patients who are stable, had prior hx of SBO, and who have SBO on ultrasound could be admitted for trial of conservative management, without obtaining CT. That doesn’t seem too unreasonable considering the acceptable cost of delayed diagnosis. If you select out the right patients, there is likely not to be too much morbidity associated with waiting to see how they do. Since this is what is done for many SBO patients already, the risk we run is that if these patients are misdiagnosed, what other diagnoses can cause problems if missed. Looking at the list of alternate diagnoses seen on CT, not many are extremely emergent, but again, this is a small population.

One problem that will eventually come up is the fact that the CT gives a lot more information than the ultrasound. Aside from evaluating for countless other pathologies, in the case of SBO, it provides anatomic information to help assess the cause and guide management. Although one ultrasound abstract has used a transition point, this is not well described and likely most people would not be able to count on finding one. Also - what about partial small bowel obstructions? What about an ileus? Some of these findings may appear similarly to an SBO on ultrasound, but the management may be different. It would be ideal to be able to differentiate those patient who would be most likely to require surgical management. If ultrasound can’t do that, it might have to be used in conjunction with other risk factors and clinical factors to help improve its utility.


Take Home Points

1. This study is small and has too many biases to take the accuracies presented at face value. We need larger, prospective studies.

2. In combination with prior studies, this study supports the concept that point of care ultrasound can accurately diagnose SBO. In this case, it showed high specificity and sensitivity in a high pre-test probability population.


Our score

2 Probes


Expert Reviewer for this Post

Minardi

Joseph Minardi, MD @jminardi21

Chief, Division of Emergency and Clinical Ultrasound, West Virginia University, Departments of Emergency Medicine and Medical Education


Reviewer's Comments

My particular practice and incorporation of this and the other POCUS SBO literature is US 1st (in pretty much all abdominal pain, vomiting), then more selective with CT imaging. In US for SBO, if I don’t see signs of it on US and I think they are low enough risk (low enough to send home), and I have no other reason I desire CT imaging, I will stop the workup. If I see signs of SBO and think conservative management will be all that is necessary (maybe in someone with history of recurrent SBO), I will stop the workup or maybe order plain films, so admitting teams may follow. If this is a new diagnosis or I think more aggressive management may be needed, I will often proceed to CT. 1. POCUS first, 2. Assess risk, 3. Anticipate next management steps. That’s just my practice incorporating this and the existing literature to the best of my ability.


Published on 11/12/18 06:00 AM
comments (0)