Although LUS cannot be considered as one of the first-choice option for patients with COVID-19 pneumonia, it has been proven to be a reliable instrument for the triage of patients when dealing with massive hospital admissions and limited CT scan capacity [Smith 2020, Lichter 2020]. LUS is relatively easy to use and is performed at bedside, minimizing the number of healthcare workers in contact with a patient. However, it can be challenging for interpretation, lung damage severity assessment and, therefore, choice of the best treatment strategy. Reliable LUS scoring protocols would be of great value as a diagnostic tool especially in emergency settings and healthcare facilities with limited access to CT scan. The only systematic review with meta-analysis on this topic concluded that diagnostic agreement between LUS and CT in the diagnosis of COVID-19 is high [Wang 2021]. The quality of the assessed evidence was considered to be low, but LUS had great potential to be an alternative to CT in emergency or intensive care setting.
In our study we developed and tested three LUS scoring protocols in patients with confirmed COVID-19. The first, 16-zone LUS protocol, was based on the scoring of B-lines, degree of interstitial changes and the size of subpleural consolidation. Two other protocols – modified 16-zone LUS NMHC and 12-zone LUS NMHC – used A-lines for scoring with B-lines having secondary importance. It was quickly recognized, that the first suggested 16-zone LUS protocol had several technical issues attributed to the scoring of chosen findings, therefore it didn’t offer either reliable scoring or convenience of execution. This protocol was used in 18 patients and in this limited sample it showed gradual increase in the score number with increase in CT severity score, which was consistent with the results for two other protocols tested in 143 patients overall.
16-zone LUS NMHC and 12-zone LUS NMHC protocols were proposed as more feasible and reliable scoring systems. Similar LUS protocols have been described in several studies [Bitar 2021, Ökmen 2021], which used ratio of vertical and horizontal artifacts, but not A-lines as used in our protocols. In the available literature, visible A-lines along with B-lines in the view of an ultrasound probe are considered the signs of severe lung damage. Meanwhile emergence of A-lines in front of B-lines is only possible when there is a certain volume of unaffected lung parenchyma, and such a depiction doesn’t correspond with any known guidelines or scoring protocols. So our hypothesis was that scoring of A-lines instead of the ratio of vertical and horizontal artifacts would be diagnostically meaningful and easier to perform than the described protocols [Bitar 2021, Ökmen 2021].
According to current evidence most of the authors tend to use 12-zone LUS protocol [Zieleskiewicz 2020, Ökmen 2021, Heldeweg 2021]. However, in our study we demonstrated that modified 16-zone LUS NMHC protocol was slightly better than 12-zone LUS NMHC. Scores from both 16-zone LUS NMHC and 12-zone LUS NMHC protocols had strong correlation with CT severity, which is complementary to similar findings in other studies [Wang 2021, Deng 2020, Rizzetto 2021, Tung-Chen 2020]. The 16-zone LUS NMHC protocol was more accurate in differentiating between <50% lung damage and > 50% lung damage as seen on CT-scan compared to 12-zone LUS NMHC protocol. That could be an argument favoring the use of 16-zone LUS NMHC protocol in emergency setting to decide which patients have worse prognosis and warrant an ICU admission.
Both protocols were borderline in differentiating poor outcome (death) from a discharge with specificity of 53% for 16-zone LUS NMHC and 55% for 12-zone LUS NMHC. Contradictory data were reported in this regard. In two observational studies [Persona 2021, Bosso 2021] higher LUS-score from 12-zone protocol was not a predictor for mortality. In contrast, Heldeweg et al. [Heldeweg 2021] demonstrated that higher LUS score from 12-zone protocol had stronger association with death and ICU stay of more than 30 days. However, in our study the score from 12-zone LUS NMHC protocol was recalculated from the 16-zone LUS NMHC protocol, which doesn’t allow for direct comparison with 12-zone protocol mentioned in these studies. Among other potential predictors, lower SPO2 at admission and higher lung damage on CT scan were associated with higher risk of death.
The main limitation of the study was its retrospective character and the fact that it was not designed to test the hypothesis of clinical equivalence rather than the correlation of LUS and CT scoring systems. Another difficulty was the use of the first 16-zone LUS protocol, that originated from the local descriptive LUS protocol for non-COVID19 pneumonia, and was abandoned early on due to several technical issues, and therefore, the sample of patients to analyze its diagnostic accuracy and correlation with CT scan was very low. On the other hand, the strength of the study was the overall sample size which was higher than in most published studies (161 patients), and the fact that two newly developed LUS scoring protocols were tested on the same patients.