Pediatric Peritonsillar Abscess - A Transcervical Protocol

By Delia Gold

Peds PTA Graphic

Impact of Transcervical Ultrasound for the Diagnosis of Pediatric Peritonsillar Abscesses on Emergency Department Performance Measures

J Ultrasound Med Nov 2019 - Pubmed Link

Take Home Points

1. An ultrasound first imaging approach to pediatric PTA has the potential to decrease LOS although this before and after study has too much risk for confounding to draw definitive conclusions.

2. Radiology-based (not point-of-care) transcervical ultrasound is feasible in this population and did not result in increased return visits or the need for follow up CT imaging.


Peritonsillar abscesses (PTAs) are somewhat common, accounting for more than 60K ED visits per year. PTA requires surgical drainage whereas peritonsillar cellulitis (PTC) requires abx only. Clinical diagnosis alone has poor performance in differentiating PTC vs PTA (sensitivity 78%, specificity 50%). CT is more accurate (sensitivity 100%, specificity 75%). Transcervical and intraoral US have specificity/sensitivity similar to CT. Intraoral US is poorly tolerated in patients with significant pain/trismus, also sterilization/cleaning is an issue. This articles examines an ultrasound-first diagnostic protocol to see if it can improve length of stay, cost, or radiation in a pediatric ED. It is not a true POCUS study since these were comprehensive sonographer performed exams, but perhaps data from this reasonably sized study with patient centered outcomes could be extrapolated.


  1. Does the adoption of transcervical US as the initial imaging choice in suspected PTA significantly reduce LOS (as opposed to using CT first)?

  2. Is the use of transcervical US as first imaging study in patients with suspected PTAs associated with a significant reduction in estimated radiation dosages and costs?


Freestanding tertiary care children’s hospital with an annual ED census of approximately 87,000 patients


  • Pediatric ED patients who had CT or US scan of neck


  • LOS information unavailable


A retrospective cohort of patients evaluated for suspected PTAs between January 2009-April 2017.

Compared ED LOS before and after implementation of transcervical US to diagnose a PTA. Collaborative protocol between EM, ENT, and radiology.

EHR was reviewed and pertinent data extracted. Patients with CT scans had estimated radiation doses, and for all patient costs were estimated. EHR also reviewed for return visits for same complaint within 2 weeks of initial visit.

Primary Outcome: Length of stay as timed by ED arrival to ED departure. Secondary outcomes were radiation dose and cost.

Assessed for confounding temporal changes in LOS by looking at all ESI 3 times through the years of the study.

Who did the ultrasounds?

This was not point-of-care ultrasound.

All scans were performed in the Division of Radiology by board-certified sonographers trained in the transcervical techniques. Images were reviewed and interpreted by pediatric radiologists.

The Scan

Linear transducer (but in adults also could use curvilinear)


Linear transducer in the submandibular region with the beam angled cephalad and posteriorly. Tonsils are immediately deep and medial to submandibular gland.

Scanned in transverse and sagittal planes with the submandibular gland serving as an acoustic window

PTA will appear as an anechoic or hypoechoic irregularly marginated area adjacent to the pharyngeal tonsil.


  • Compare both sides for subtle asymmetry. Use dual screen if you want.

  • The on screen image can be flipped upside-down to align more with anatomical orientation (since your transducer is facing up).

  • It is nice to identify the carotid artery with color before any drainage procedure.

Learn how to do a Submandibular Approach for PTA from MGH Ultrasound!

Check out Peritonsillar Abscess Pathology on the POCUS Atlas!



387 eligible patients → 101 evaluated with CT and 286 evaluated with US

  • None of the patients who had US scans required a subsequent CT scan

  • Similar rates of CT and US studies with positive results

Primary Outcome - Length of Stay in Emergency Department

CT group = 426 +/- 171 minutes

US group = 347 +/- 145 minutes.

Absolute difference of 79 minutes (95% CI 44-113 minutes)

✳ Difference in time was only statistically significant in those with negative results (CT group mean LOS 115 minutes longer than US group). When there were positive results mean difference was 12 minutes which was NOT statistically significant.

Secondary Outcomes


No patients required CT after US. The radiation was estimated per CT scan.

CT 3.9 mSv

US 0.0 mSv


Estimated difference in cost was calculated based on the mean billed cost per study. For CT this was $2846 and for US this was $1208. This was then multiplied by the total number of patients seen in the study to represent cost different between an all-CT and an all-US scenario.

CT $1,101,402

US $467,496

Other Findings

There was not a significant difference in rate of return visits within 2 weeks of index visit (5.9% in CT group vs. 8.0% in US group).

Significantly more US were ordered (mean difference 4.4 studies per month).


The main limitation is that this is a retrospective pre and post study. This leaves a lot of vulnerability to cofounders. The CT group was January 2010-April 2013 vs. US group May 2013-April 2017. They tried to assess any changes in LOS by stating that the average LOS for ESI level 3 patients was actually longer in the post phase of the study. Some of the cofounding could have been addressed by collecting more data. They only compared CT before protocol to ultrasound after protocol? Why not just compare CT to ultrasound? Furthermore, they only studied patients who received CT or US. What if a patient had had neither? The study could have missed obvious cases not requiring imaging or more subtle presentations in which imaging was not considered.

There was no comparison of the acuity or severity of illness of the two groups. Maybe the CT group was sicker or required sedation? Many other factors can affect LOS that were not reported here.

Uneven final numbers (CT 101 vs. US first 286). 34 patients had imaging performed for suspected PTA but LOS information was not available so they were excluded from study. These 34 patients would have represented 8.8% of the total population enrolled which is not insignificant.

This is not POCUS. We cannot necessarily extrapolate this to POCUS since this is fairly novel and many do not have experience with this application of ultrasound.


Difference in LOS seemed to be only when in the subset with negative findings on imaging. This is likely due to other systemic factors present when patient has the diagnosis (consult time, abx, labs, admission).

Additionally, after implementation of US first protocol, there were significantly more US ordered as opposed to CTs. There were 4.4 more ultrasounds/month compared to CT which is not surprising based on fondness for US in general in pediatrics. This means that these populations could be different in terms of their risk for PTA (although percentage of positives studies was similar). Having a more appealing imaging option is a double edged sword – if a CT is the only option, in theory the PEM is going to debate the need for the study more (radiation/IV/cost), as opposed to an US (which is perceived as harmless) could have led to overall increase in imaging overall and hence overall cost.

Take Home Points

1. An ultrasound first imaging approach to pediatric PTA has the potential to decrease LOS although this before and after study has too much risk for confounding to draw definitive conclusions.

2. Radiology-based (not point-of-care) transcervical ultrasound is feasible in this population and did not result in increased return visits or the need for follow up CT imaging.

More Great FOAMed on this Topic

Ultrasound Podcast - Telescopic PTA with Brooke Hensley and Andrew Liteplo

Our score

2 Probes

Expert Reviewer for this Post


Jeremy S. Boyd, MD, FACEP @geek_md

Living and working in Nashville, TN as Ultrasound Fellowship Director at Vanderbilt University Medical Center and another hospital that starts with a V.

Reviewer's Comments

Regarding the article, I really appreciate that the study’s outcomes of interest are clinically oriented ED measures. It’s important for clinical ultrasound that we continue to ask these kinds of questions in our research—morbidity/mortality is a great outcome, but often a herculean reach in any POCUS study design, particularly given the fact that POCUS is a diagnostic tool, not a treatment. I applaud the authors for choosing clinically relevant outcomes where the odds of finding a signal amongst the noise are within reach—particularly in a relatively novel POCUS application within pediatric emergency medicine.

That being said, the reviewer is on point—there are too many confounders here to clearly distinguish whether an ultrasound first approach is decidedly better. That being said, the study question is a clinically important one, and this study sets the stage for a prospective study design that could encompass both radiology and clinician performed ultrasounds in conjunction with CT diagnosis of PTA. The fact that no patients received CT imaging after having had ultrasound performed is almost too good to believe—if it were to hold true in a prospective study (even if mostly true) that would be a huge win for POCUS and for patients!

Bottom line: Great take on an interesting albeit imperfect study. And props to the study authors for planting the seeds for a great prospective POCUS study in the future!

Cite this post as

Delia Gold. Pediatric Peritonsillar Abscess - A Transcervical Protocol. Ultrasound G.E.L. Podcast Blog. Published on May 11, 2020. Accessed on December 05, 2020. Available at
Published on 05/11/20 05:00 AM
comments (0)