We thank you and readers for the great opportunity to extend the discussion on our study.1 We answer point-by-point to the three letters.
To Dr. Foti et al.
First, the “B-pattern” is not a new terminology. It is a recommended terminology agreed on by the main international experts and represents a standard definition worldwide.2 Regarding the “A-pattern,” there is no confusion with A-lines that are meaningless and never considered for hemodynamic assessment.
Second, regarding the number of cases, we here confirm the conclusions and also the limitations already reported in our article.1 We reiterate that the correlation with pulmonary artery occlusion pressure (PAOP) was not satisfactory, which represents the first interesting finding of our study and is reflected even in the title of our article. Indeed, pulmonary congestion and PAOP may be independent, which opens new perspectives about the usefulness of B-lines and is worthy of further investigation in future studies. Regarding extravascular lung water (EVLW), the combination of A-pattern and EVLW greater than 10 ml/Kg was found in 1 of 32 patients and not 3 as is superficially reported in the letter. This unique patient had an EVLW level of only 1 point above the threshold (11 ml/kg). However, the cutoff chosen may be crucial because there is no perfect consensus on the abnormal value of EVLW. Extending the study population and testing different cutoff values may help to assess the real relation between B-lines and EVLW calculated by the PiCCO technology. Thus, we reiterate that the absence of B-lines may safely guide fluid resuscitation in critically ill patients, whereas the possible appearance of B-lines during fluid administration demands discontinuance. This principle, more authoritatively evidenced by previous studies, is further supported by our data.3,4 However, 4 of 32 cases showed a combination of diffuse B-lines and EVLW less than 10 ml/kg. This is not surprising because the phenomenon of B-lines is not only linked to pulmonary congestion but also to acute respiratory distress syndrome and infections, which represent more complex pulmonary conditions. Avoiding fluid overload in these patients is safe in any case. When the B-pattern is detected, our suggestion is to avoid or consider more carefully the fluid administration by relying on adjunctive tools. Of course, fluids are better guided once invasive monitoring is established, which is clearly reported and continuously repeated in clear words in our article. We leave to the judicious opinion of Foti et al. but especially of anyone else who really faces complex emergency situations at bedside, whether this potential is useful or not. Indeed, B-lines cannot be used to predict fluid responsiveness but only the lung tolerance to fluid administration. However, even when we have the possibility to use more advanced tools for hemodynamic monitoring, like usually happens once the patient is admitted to the intensive care unit (ICU), B-lines assessment still maintains its importance because it is the only bedside tool that allows a safe direct insight into the lung.5
Finally, we do not understand why Foti et al. extrapolated some sentences from our article and interposed their words to change the original meaning. We did not write that “even” the combination with left ventricular ejection fraction (LVEF) did not improve the accuracy of lung ultrasound. Foti et al. will understand that if they add the word “even,” the meaning of our sentence totally changes. What is true is that, in the EVLW group, LVEF did not increase the performance of lung ultrasound, simply because it was sufficiently accurate by itself.
To Dr. Melillo
First, our article does not “advocate the usefulness of B-line counts” because we did not count B-lines, which is clearly explained in the method.1 We used a very basic and simple dichotomous method.
Second, Dr. Melillo writes that “experience dictates” that “B-lines are poorly reproducible.” Exactly the opposite is true. A large body of scientific evidence dictates that B-lines are highly reproducible and easy-to-learn by operators with different skill and expertise. Agricola et al. obtained intraobserver and interobserver variability of 3.1 and 4.4% on 280 scans performed by two operators.4 Boussuges et al.6 obtained interobserver variability less than 5% on 1,890 scans performed by two operators. Fagenholz et al.7 obtained “excellent agreement” on 308 scans. Jambrik et al.8 obtained intraobserver and interobserver variability of 5.1 and 7.4% on 560 scans performed by two operators. Liteplo et al.9 obtained interobserver reliability Cohen κ index of 0.82 on 800 scans performed by several operators with different skill and expertise. Mallamaci et al.10 obtained interobserver concordance index of 0.96 on 672 scans. Volpicelli et al.11 obtained interobserver variability of 4.9% on 192 scans performed by five operators. Other more recent studies, published after the consensus conference (from 2012 onward) and using more standardized techniques, obtained even better results.2 In a recent very large multicenter study, Pivetta et al.12 obtained a κ statistic for agreement of 0.94 on 1,200 scans performed by several operators and then reviewed by one expert and two residents with limited training. In the same study, intraobserver agreement was 0.97 for the expert operator and 0.92 for the physicians with limited training.12 Regarding the assessment of LVEF, the eyeball visual estimation represents an easy-to-learn technique, highly reproducible, and highly accurate in comparison with more advanced ultrasound and also invasive techniques.13–20 Again, exactly the opposite of what Dr. Melillo questions in the letter.
Third, we insist that the wide clinical spectrum of our study is a strength. Observing only a selected population without the possibility of extending the conclusions to the common emergency clinical practice would indeed represent a limitation. The early hemodynamic assessment is a common task in any patient, including all the conditions listed by Dr. Melillo. It is not clear from the letter what a clinician should do while waiting for more definitive monitoring methods to be established in these cases. For instance, we wonder why Dr. Melillo does not question with the same vehemence the high variability and low accuracy of other bedside tools, like chest radiography and physical examination. Does Dr. Melillo know the data of variability and accuracy of chest radiography and the physical signs for predicting pulmonary congestion? Moreover, contrary to what Dr. Melillo questions, lung and cardiac ultrasound have been largely “tried and tested.”
To Dr. Cipriani et al.
First, the letter is based on an incorrect premise. We did not consider the combination of LVEF and lung ultrasound as the “most important finding of the study,” whereas we agree that LVEF “cannot be chosen as a target tool to assess hemodynamic state.” The rationale for combining the assessment of LVEF was to improve the prediction of B-lines for high PAOP. It is well known that B-lines are reliable signs of loss of lung aeration but do not differentiate acute respiratory distress syndrome, a condition with low PAOP, from cardiogenic edema, characterized by high PAOP.5 Indeed, although 3 of 11 patients in our study and 22 of 35 in a previous study showed a combination of B-pattern and low PAOP, none of the 8 patients with B-pattern and impaired LVEF had low PAOP.1,3 Also, a normal LVEF increased the prediction of the A-pattern for low PAOP. This latter finding intuitively means that patients with impaired LVEF may be more easily characterized by a separation between hemodynamic and pulmonary congestion, thus reducing the reliability of B-lines standing alone. We agree that detailing the diagnoses and chronic/acute characteristics of these patients is crucial, and this is why we added tables 3 and 4.1 We also think that this topic is worthy of further investigations, which may only be accomplished by extending the population and using more advanced cardiac ultrasound techniques. The take-home message from our study and from the previous significant literature is that although B-lines are enough to predict pulmonary congestion, B-lines should be combined with LVEF to improve the prediction of PAOP.
Second, we do not know whether Cipriani et al. are used to caring for critically ill patients in the ICU, as their institution appears to be devoted to internal and geriatric medicine. However, their questions on the supine or seated position and dyspnoic situation of ventilated patients in the ICU seem quite strange and unusual. Moreover, it is again false and perhaps due to incomplete reading that we did not comment on the combination of A-pattern and high PAOP. We clearly wrote in the article that this finding, although originally unexpected, may be considered the most important finding of our study, and we fully discussed and referenced our explanation based on the difference between hemodynamic and pulmonary congestion.
Third, it is disarming how Cipriani et al., in common with the other authors of this group of letters, do not consider the previous literature on the topic, when they claim in their letter “…having no study been performed to actually validate…” Please read at least some of the previous literature reported in this reference section.3,4,21
Competing Interests
The authors declare no competing interests.