Wells score with POCUS improves accuracy and could decrease advanced imaging orders
Nazerian P, Volpicelli G, Gigli C, et al; Ultrasound Wells Study Group. Diagnostic Performance of Wells Score Combined With Point-of-care Lung and Venous Ultrasound in Suspected Pulmonary Embolism. Acad Emerg Med. 2017 Mar;24(3):270-280.
Patients presenting with signs and symptoms that are concerning for pulmonary embolism (PE) are typically risk stratified - often using the Wells score. Those who are at low risk by the Wells score undergo D-dimer testing as the initial strategy, while those with a high-risk score go directly to advanced imaging, usually CT pulmonary angiography (CTPA). The authors hypothesize that the Wells score can be improved by integrating the use of point-of-care ultrasound (POCUS). Their primary objective was comparing the traditional Wells score to the ultrasound Wells score.
The authors proposed changing two elements of the Wells score. The “signs and symptoms of DVT” was changed to “venous ultrasound positive for DVT” by performing a POCUS compression venous ultrasound on both legs. If a DVT was found, then 3 points were added, while 0 points were added if no PE was found. The “PE most likely diagnosis” was changed to “alternative diagnosis less likely than PE after lung ultrasound”. This was done by performing a lung POCUS. If a pulmonary infarct was found, then 3 points were added. If no pulmonary infarcts were found AND an alternative diagnosis (e.g. Pneumonia) was found, 0 points were found. In the case of a normal lung ultrasound, points were assigned based on the conventional Wells score. The rest of the Wells score was interpreted as usual.
The study was a multicenter prospective observational study in four Italian EDs. Patients were enrolled if the attending physician was pursuing the diagnosis of PE. A D-dimer was drawn in all included patients. The attending physician assigned points for all the remained Wells criteria. All included patients had a venous and lung POCUS performed by either an attending or resident with at least 6 months of POCUS experience. A Wells score (WS) and ultrasound Wells score (USWS) was calculated in all patients.
The authors enrolled 446 patients, in whom 69% received either a CTPA or a VQ scan. The PE prevalence was 28%. The use of USWS increased the sensitivity of the score from 57.6% to 69.6% and increased the specificity from 68.2% to 88.2%. The use of lung and venous ultrasound without the use of D-dimer was not sufficiently sensitive to exclude PE. The authors note that using the USWS was 5% more efficient than using the WS. When adding in a negative D-dimer, the USWS had a sensitivity of 99.1% compared to 98.2% for WS, while the specificity for USWS was 44.6% compared to 36.8% for the WS. The authors state that the use of USWS would have decreased the number of CTPAs performed, from 50.2% to 27.2%, as many patients with a high WS would have been reclassified as low if using the USWS.
It is no surprise that adjusting the WS using POCUS would lead to a more accurate score. I commend the authors for studying it in what is the first study of likely more on this topic. Emergency physicians are always looking for ways to decrease the number of imaging studies. Physicians performing the POCUS had various levels of training, making it applicable even when physicians without extensive ultrasound experience are available. While the USWS is more accurate, it was only 5% more accurate and it did not perform so well as to not require a D-dimer. Additionally, performing the USWS requires more time (7 +/- 3 minutes). The slight increase in time must be balanced with the only slight increase in efficiency. On the other hand, if it truly does decrease CTPA use by 23%, then increased time would be justified.
As the authors note, they did not directly study the accuracy of a diagnostic strategy using USWS plus D-dimer and therefore this study can only be hypothesis generating. I found this study intriguing and look forward to future research on the subject.
Subxiphoid view is best when assessing IVC collapsibility index
IV fluid resuscitation is a cornerstone of resuscitation. We know that under- or over-administration of fluids can lead to worsening morbidity and mortality. The assessment of the inferior vena cava (IVC) using ultrasound is well described and incorporated into many POCUS assessments. In addition, multiple ways to assess the IVC have been described, of which minimal data comparing each has been assessed. This study aimed to compare three IVC acquisition techniques. The views assessed are the subxiphoid transabdominal long axis view, transabdominal short axis view just inferior to the hepatic veins and the right lateral transabdominal coronal long axis view.
Three POCUS experts (2 attending staff and one ultrasound fellow) assessed the IVC in each view using both B- and M-mode. The IVC collapsibility index was calculated. The POCUS experts assessed 39 volunteers, totaling 351 measurements. A well described protocol is laid out in the paper. Statistical analysis included inter-rater reliability, and intraclass correlation coefficients.
The study found that the subxiphoid long axis view of the IVC produced the highest inter-class correlation. The short-axis and lateral transabdominal views were less reliable. Post hoc analysis revealed ultrasonographers had difficulty in obtaining consistent images. The study also states that it adds to the growing literature that the IVC assessment should consist of just B-mode, and there is no need to introduce M-mode scanning.
This was a well-done study that provided an answer to an area of IVC POCUS that was not fully explored. My main critique is not as much about the study itself, but the reality that the IVC may not be all it’s cracked up to be with regards to volume assessment. Let the debate ensue!
Cardiac Journal Summary
Hypertrophic cardiomyopathy (HCM) is a potentially asymptomatic disease that often presents for the first (and only) time as sudden cardiac death during strenuous physical exertion. The American Heart Association (AHA) reports that 36% of cases of sudden cardiac death (SCD) in athletes can be attributed to HCM.1 The true incidence of genetic HCM is unknown due to this abnormality often remaining clinically silent. The natural history of this disease makes it ripe for screening, however due to its low prevalence, the AHA does not support a nationwide screening program in student-athletes. They do however endorse targeted screening through individual small-scale projects in focused geographic regions.2 Pre-participation screening for student-athletes commonly entails a focused personal and family history as well as a physical exam, with subjects that are found to have abnormalities referred for electrocardiogram (ECG) and/or echocardiogram testing. In the June edition of the Journal of Ultrasound in Medicine, Fox et al. report the findings of a targeted HCM echo-based screening program utilizing medical students as the primary members of the research team.3
What They Did?
They prospectively enrolled student-athletes from 12 area high schools & 3 area colleges near the campus of University of California-Irvine from May 2012-August 2013. Student-athletes between the ages of 14-30 years old that were competing in any school-sponsored sporting activity were eligible for participation. Eligible participants were recruited through the distribution of IRB-approved information pamphlets as well as consent forms and HIPAA forms by their school administrators. The research team consisted primarily of medical students in their first and second year of school. All medical students at the institution undergo 8h of didactic and 8h of hands-on point-of-care ultrasound (POCUS) training during gross anatomy in their 1st year and another 8h of didactic and hands-on POCUS during pathology in their 2nd year. In addition to this, students that participated in the research teams reviewed a slide presentation detailing cardiac ultrasound for the project and attended a 2h hands-on training workshop led by trained cardiac sonographers. Lastly a competency examination needed to be passed by students before they were able to enroll subjects. In all 35 different medical students performed all study echocardiograms using Sonosite S-FAST ultrasound machines (FUJIFILM SonoSite Inc, Bothell, WA).
The students worked in teams of 2 for each POCUS examination, one student acquired the images and the other operated the machine. Required images for each exam included a parasternal long axis (PLAX) clip, a parasternal short axis (PSAX) clip and lastly a PSAX clip utilizing M-mode to measure the interventricular septum (IVS) thickness, the left ventricular chamber diameter and the left ventricular posterior wall (LVPW) thickness. All study echocardiograms were reviewed by a board-certified pediatric cardiologist who was blinded to all participant information. Participants with an IVS:LVPW thickness ratio greater than 1.25 or with LVPW thickness greater than 12mm were considered abnormal and underwent repeat echocardiogram by the pediatric cardiologist.
What They Found?
A total of 2402 student-athletes underwent POCUS exams as part of the HCM screening project, 70 (2.9%) of who were excluded because of poor image quality. This left 2332 participants that were enrolled, however another 61 were excluded because the POCUS exams lacked demographic information, thus leaving the total number of participants who had their data analyzed at 2271. A total 137 of these participants (5.8%) were found to have an IVS:LVPW thickness ratio greater than 1.25. All 137 of these students underwent repeat scanning by the pediatric cardiologist with 7 of these participants (4.86%) found to have HCM. The authors report a sensitivity of 100% and a positive predictive value of 4.86%. Of the patients with HCM, 6 were male (85.7%), 4 (57.1%) were black, 2 (28.6%) were Asian and 1 student with HCM did not identify her race.
What This Means?
The authors report success in one of their primary objectives of determining whether medical students with a small amount of targeted training could successfully screen for HCM. They support this by pointing towards acceptable image quality and enrollment in 94.5% of the participants (2271/2402) and through a low false positive rate of 5.6% which compares favorably with other HCM screening protocols such as the European Society of Cardiology’s and the Seattle Criteria.4,5 I would caution the authors somewhat regarding the use of the low false positive rate as a marker of the program’s success because this rate reflects more on the very low prevalence of the condition (HCM) in this population (7/2271=0.3%) than it does on the characteristics of the test (POCUS).
The results of this study are also subject to the partial verification bias, in that only participants with an abnormal screen received the confirmatory test, so the true rate of HCM in this population is unknown since 2134 subjects did not receive confirmatory testing. Also, the authors chose a very conservative threshold for IVS:LVPW thickness of 1.25 which likely resulted in such a large proportion of the participants with an abnormal screen not going on to be diagnosed with HCM. This was an intentional decision on the authors part in order to maximize the sensitivity of the screening test while sacrificing specificity most likely because the subjects were being scanned by minimally trained medical students.
The primary component of this program’s success is that 35 medical students were able to perform adequate POCUS exams to screen for HCM in a relatively large number of student-athletes (2271). This is a massive undertaking and the authors should be lauded for being able to pull this off. Another element that is described by the authors as a sign of the program’s success is the limited time it took each medical student team to complete the POCUS exam on each student-athlete (on average 5 minutes). Given the increasing frequency with which medical schools are incorporating POCUS into their curriculum, a project such as this should give leaders in our specialty hope and inspiration that by utilizing a cohort of highly motivated medical students with targeted training above and beyond the common POCUS curriculum that big ideas such as this can be brought to fruition. Student-run clinics are very common across the country (as the authors state in their Discussion section), and may be an ideal setting for further studies as to the utility of POCUS exams to screen for various conditions. We came away very inspired after reading this manuscript and we hope that the readers of this month’s cardiac journal summary are similarly inspired to do great things.