• 
• 

Three Part Question

In [critically ill patients presenting to the emergency department with sepsis requiring fluid resuscitation], does [bedside lung ultrasound or b-lines assessment] be used to determine the [pulmonary fluid status]?

Clinical Scenario

A 68 year-old male is brought to the emergency department with tachypnea (32 breaths per minute), tachycardia and a core temperature of 39.2°C. He complains of a progressive back pain since 3 days and recently noticed hematuria. In the emergency department, his systemic arterial blood pressure is 80/56 mm Hg and his heart rate 136 beats per minute. You suspect a severe sepsis from an acute pyelonephritis. To restore the blood pressure, you administer a first bolus of intravenous crystalloid. To guide your fluid therapy, you wonder if bedside lung ultrasound will help you assess the interstitial fluid status of your patient to avoid over-hydration.

Search Strategy

Search made on november 18th 2016
A) No BestBETs or critical appraisals were found on this topic.
B) The website www.clinicaltrials.gov was searched for ongoing trials on the subject. No trial was found.
C) The Cochrane Library was searched for reviews on the subject. No reviews were found.
D) MEDLINE using the PubMed interface (1950 – 2016/11/18)
#1 – 'lung' OR 'b-lines' // 761, 723 papers
#2 – 'echography' OR 'ultrasound' OR 'ultrasonography' // 507, 513 papers
#3 – #1 AND #2 // 14,515 papers
#4 – 'fluid resuscitation' OR 'fluid assessment' OR 'fluid therapy' OR 'volume repletion' OR 'bolus fluid' // 108, 125 papers
#5 – #3 AND #4 // 325 papers
#6 – Limit to Humans // 272 papers

After reviewing the articles, 4 relevant papers were found.

In addition, reference lists of relevant papers were checked for potential studies. One study protocol not yet recruiting participants was found: Long, E. et al., 2016, Australia [1] This prospective observational study will explore the effect of fluid bolus on echocardiogram, IVC and lung ultrasound in children with clinically diagnosed sepsis. Measurements will be taken before the fluid bolus and repeated 5 and 60 min after. The extravascular lung water will be scored for the presence of B-lines to calculate a lung ultrasound score.

E) EMBASE using the Elsevier interface (1966 – 2016/11/18)

#1 – 'lung' OR 'b-lines' // 1, 413, 328 papers
#2 – 'echography' OR 'ultrasound' OR 'ultrasonography' // 613,821 papers
#3 – #1 AND #2 // 34, 640 papers
#4 – 'fluid resuscitation' OR 'fluid assessment' OR 'fluid therapy' OR 'volume repletion' OR 'bolus fluid' // 28, 911 papers
#5 – #3 AND #4 // 247 papers
#6 – #3 limit to humans // 233 papers

After reviewing the articles, 4 relevant papers were found, but 2 were duplicates from those found on Medline, adding one more relevant paper. In addition, reference lists of relevant papers were checked for potential studies. No new studies were found.

Search Outcome

356 papers were found (excluding duplicates). After reviewing the articles, 352 papers were excluded as they were found irrelevant to the question, of insufficient quality (consisting mostly of expert article or review articles) or because they were written neither in english nor in french. Papers examining echocardiography (VCI or LV contractibility) for the fluid resuscitation or using ultrasound for fluid status before and after hemodialysis were also excluded. A total of 4 relevant papers were retained in this review.

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Enghard et al 2015 Germany	All ventilated patients 18 years or older admitted to medical ICUs for various diagnoses in Berlin (n=50).	Prospective observational study	Four-sector lung ultrasound for EVLW assessment	Close correlation between the presence and extend of B lines with (EVLW) (Spearman’s r = 0.72, P < 0.0001)	ICU patients admitted, ventilated. No dynamic measure to fluid therapy has been evaluated. Relatively small study Single center Gold standard used (Transpulmonary thermodilution to measure (EVLW)) and the cutoff values remain arbitrary in the literature. No info on the delay between the lung ultrasound and the gold standard measure. Heterogeneity of the patient diagnosis. (Only 34% 17/50 had sepsis) No information on ultrasound operators, their training. No documentation of inter-observer reliability (kappa score).
Theerawit et al, 2014 Thailand	Patients ≥18 years with the presence of shock and requiring fluid therapy, admitted to medical ICU of a university hospital (n=20)	Prospective observational study.	Total B line score (TBS) and total fluid volume at admission and 48h after the fluid resuscitation in 23 lung points of assessment	ΔB-lines along anterior axillary line (AAL) high correlation to ΔTBS (Pearson r = 0.90, P < 0.01)	ICU patients admitted, some ventilated. Only anterior axillary line had inverse correlation to PaO2/FiO2 ratio, which is secondly correlated to ΔTBS. The long 48h delay between the initial and second lung ultrasound. Small sample size. No actual EVLW gold standard measurement Partial blind study 2 patients had pneumonia and one patient had ARDS, which might have influenced the Δ PaO2/FiO2 ratio and Δ B-lines at all area. Only 20
			PaO2/FiO2 ratio (P/F ratio)	ΔB-lines along anterior axillary line (AAL) inverse correlation to PaO2/FiO2 (Pearson r = -0.704, P < 0.06) Increase of B-lines ≥ 10 was related to the decrease of PaO2/FiO2 ratio.
Lichtenstein, DA 2013 France	Patients in acute circulatory failure (distributive shock, i.e. septic shock prone to fluid resuscitation).	Review article Expert opinion	Detection of pulmonary edema as an early marker of fluid overload, called FALLS-endpoint.	Absence of multiple B-lines excludes pulmonary oedema and define FALLS-responder.	Broad population patients Qualitative fluid administration guide. No measure. Expert opinion No validation of the FALLS-protocol. No information on ultrasound operators, their training.
			Detection of pulmonary edema as an early marker of fluid overload, called FALLS-endpoint.	obstructive shock, A-profile can receive additional fluid. A change from A to B-lines (3 being visible) will demonstrate interstitial syndrome at an early sign of pulmonary oedema.
Lichtenstein et al 2009 France	Critically ill patients with a PAC (pulmonary artery catheter) in UCI for various diagnoses n=102	Prospective observational study.	A-lines predominance or B-lines predominance by lung ultrasound for detecting anterior interstitial oedema.	At PAOP < 13mmHg A predominance had 90% specificity and 67% specificity, 91% PPV.	Broad population patients ICU patients admitted, all ventilated. No dynamic measure to fluid therapy has been evaluated. Single centre PAOP for pulmonary fluid assessment has declined in ICU since its value is being debated in the literature. Heterogeneity of the patient diagnosis. (Only 24% 24/102 had sepsis) No information on necessary training needed to perform this lung ultrasound.
			A-lines predominance or B-lines predominance by lung ultrasound for detecting anterior interstitial oedema.	At PAOP < 18 mmHg A predominance had 93% specificity and 50% specificity, 97% PPV. A predominance suggests that fluid may be given without concern for the development of hydrostatic pulmonary oedema.

Comment(s)

Fluid resuscitation for sepsis is shown to improve survival in adults with septic shock, although the harms from cumulative positive fluid balance have become apparent. Currently, the most accurate method to guide fluid administration decisions uses dynamic measures that estimate the change in cardiac output in response to a fluid bolus. How to assess the end point where the patient has received optimal fluid therapy while the signs of circulatory failure persist remains unclear. Lung ultrasound has gained attention as a rapid and non-invasive tool to detect a shift from dry lung to B-lines, an artefact indicating interstitial oedema or alveolar oedema (confluent B-lines), which prevent fluid overload. Regarding this emerging clinical practice, this review aims to evaluate if bedside lung ultrasound adequately assesses pulmonary fluid status in critically ill patient with sepsis requiring fluid resuscitation. From this review, there is missing quality evidence in the current literature guiding the fluid therapy in clinical context. The search strategy elicited only three prospective observational studies with low quality evidence. A consistent difficulty in these studies is the lack of dynamic measures before and after bolus to guide the fluid therapy. Also, various methods are currently used in the literature for the pulmonary fluid assessment without no consensus on the gold standard. The pulse contour cardiac output (PiCCO) invasive transpulmonary thermo-dilution method to measure (EVLW) index is described as the reference method in ICUs although another study used the pulmonary artery occlusion pressure (PAOP) as a marker of lung filtration pressure. More over, patients from those studies differ from our clinical question as they were mostly ventilated and admitted in intensive care unit with heterogeneity of diagnosis. The study by Enghard, P, et al, 2012 obtained a significant correlation between the presence and extent of B-lines with increasing EVLW index. Theerawit P, et al, 2014 obtain an inverse correlation to PaO2/FiO2 ratio with B-lines along anterior axillary line access by lung ultrasound.1 reported a VPP of 97% between A-profile (absence of B-lines) by lung ultrasound and a PAOP <18 mm Hg suggesting that fluid may be given without initial concern for the development of hydrostatic pulmonary oedema on critically ill patients. Those three papers are the only studies presenting quantitative measurements with their methodological limitations and partially responding to our clinical question. Finally, Lichtenstein, D, 2013 an expert opinion paper presents is protocol that supports fluid resuscitation guided by lung ultrasound. After excluding cardiogenic shock and obstructive shock, the FALLS-protocol indicates that patients with an A-profile can receive additional fluid. If the dominant mechanism is vasoplegia, fluid therapy will not improve the shock, and eventually a change from A-lines to vertical B-lines (three being visible) will demonstrate interstitial syndrome, an early sign of interstitial oedema in fluid overload. This protocol has not been validated.

Editor Comment

AAL, anterior axillary line; ARDS, acute respiratory distress syndrome; PPV, positive prediction value; TBS, Total B line scoreTBS

Clinical Bottom Line

In patients presenting to the emergency department with severe sepsis requiring fluid resuscitation the use of lung ultrasound for determining the pulmonary fluid status remains unknown. The current data supporting this practice is missing. A coming prospective observational study will better document this clinical question by measuring the effect of fluid bolus with lung ultrasound in children diagnosed with sepsis.

References

1. Enghard P , Rademacher S , Nee J , et al . Simplified lung ultrasound protocol shows excellent prediction of extravascular lung water in ventilated intensive care patients, Crit Care 2015;19:36.
2. Theerawit P , Sutherasan Y , Kiatboonsri S , et al . Transthoracic ultrasound assessment of B-lines for identifying the increment of extravascular lung water in shock patients requiring fluid resuscitation. Indian J Crit Care Med 2014;18:195–9.
3. Lichtenstein, DAF. FALLS (Fluid Administration Limited by Lung Sonography)-protocol: lung ultrasound in hemodynamic assessment of shock. Heart, Lung and Vessels 2013;5:142–7.
4. Lichtenstein DA , Mezière GA , Lagoueyte JF , et al . A-lines and B-lines, Lung ultrasound as a bedside tool for predicting pulmonary Artery Occlusion Pressure in the critically ill Chest 2009;136:1014–20.