Ultrasound G.E.L. – Inferior Vena Cava Collapsibility for Fluid Responsiveness
- Sep 27th, 2019
- Michael Prats
- categories:
Originally published on Ultrasound G.E.L. on 7/31/17 – Visit HERE to listen to accompanying PODCAST! Reposted with permission.
Follow Dr. Michael Prats, MD (@PratsEM), Dr. Creagh Bougler, MD (@CreaghB), and Dr. Jacob Avila, MD (@UltrasoundMD) from Ultrasound G.E.L. team!
Inferior vena cava collapsibility detects fluid responsiveness among spontaneously breathing critically-ill patients
Journal of Critical Care May 2017 (Epub) – Pubmed Link
Take Home Points
1. In spontaneously breathing patients, an IVC collapsibility index >25% may have some predictability of volume responsiveness, but this data is not sturdy enough to rely on it definitely.
2. This study did not show improved predictability with dynamic measurements or passive leg raise which is contrary to prior literature.
Background
Every patient has an optimal balance where they have just the right amount of volume in their system. It can often be difficult to determine where patients are in that balance based on physical exam and history alone. You don’t want to give a patient too little volume such that their body is not functioning properly, but you also don’t want to drown them by pouring too much fluids through their veins. This especially comes into play with critically-ill patients. Often times, these patients are hypotensive and the question is how much fluids to give. There are certainly invasive and expensive ways that you can try to find this out, but in order to make quick bedside decisions – it would be ideal if ultrasound could help out. We know that ultrasound of the IVC is pretty good in mechanically ventilated patients with regulated tidal volumes, but IVC measurements for fluid responsiveness in the spontaneously breathing has gone back and forth. First it was thought to be useful, then useless, maybe only useful at the extremes – well this paper seeks to figure out exactly how it might be used.
For a good understanding of the terms used and the important distinction between volume status, volume responsiveness, and volume tolerance – check out this great post by @thinkingcc
Volume status vs Volume responsiveness vs Volume tolerance
Questions
Can POCUS of the IVC detect fluid responsiveness in spontaneously breathing critically ill patients?
What is the optimum cutoff for the collapsibility of the IVC?
Does incorporating a passive leg raise help in determining fluid responsiveness?
Population
Enrolled from August 2014 to July 2016 at two adult academic hospitals in the US
Inclusion
Acute circulatory failures as defined by
A) systolic blood pressure <90 mhg, or MAP <65 for ≥30 minutes
B) decreased urine output (<0.5ml/kg/h)
C) tachycardia >120 bpm ≥30 minutes
D) serum markers suggesting hypoperfusion (pH <7.3 or lactic acid >2 meq/L)
Exclusion
- Primary traumatic, cardiogenic, obstructive, or neurogenic shock (not clear how this was determined)
- Age <18 yo
- Incarceration
- Pregnancy
- Hospitalization for >36 hours [prior to enrollment]
- If receiving NIPPV
- If clinical team thought they had active pulmonary edema
- If clinical team believed that IVF might “pose clinical risk”
Design
Prospective, observational, convenience sample
IVC ultrasound was compared to NICOM (non-invasive measurement of cardiac index) to determine correlation
Fluid responsiveness defined as ≥10% increase in cardiac index as measured by NICOM
90% power estimate determined sample size of 124 patients to detect difference in AUC (area under the receiver operating characteristic curve) of 0.88 and AUC of 0.70
Clinical and demographic characteristics summarized and assessed for significant differences
Intra and interrater reliability calculated from baseline IVC calculations.
Examined baseline IVC measurement compared to change in cardiac index
Also looked at IVC inspiratory and expiratory diameter or change in cIVC after PLR or 500 mL bolus
First checked fluid responsiveness of cIVC and PLR separately then checked if a 5% change in cIVC after PLR would identify fluid responders
Intervention
Following enrollment the patients had NICOM device leads placed which records cardiac index at 1 minute intervals
Placed supine for three minute NICOM calibration period
Then two baseline 10 second videos of IVC, 1 minute apart (#1 and 2)
Then 3 minutes passive leg raise (I believe that classic straight leg raise goes from sitting to supine with legs up)
Then 1 more IVC video (#3)
Then returned to supine for 3 minutes minimum.
Then 500 mL normal saline bolus administered through pressure bag through largest gauge IV
After bolus completed, single IVC video (#4)
Vasopressors (if patient was requiring) were held at constant rate throughout study
During review, images were frozen during maximum expiratory and minimum inspiratory diameter. Measured 3 cm caudal to IVC right atrium junction.
Calculated cIVC (collapsibility of IVC or caval index)
cIVC = (IVC expiratory diameter – IVC inspiratory diameter)/IVC expiratory diameter
Reviewer were blinded to NICOM results
Who did the ultrasounds?
3 study physicians (from departments of emergency medicine and pulmonary critical care medicine)
Training included POCUS (but does not specify to what degree, only mentioning they did not have ultrasound fellowship training)
The Scan
Phased array probe (1-5 MHz – actually says Hz in this paper – really low!)
Subcostal long axis view
Study participants not asked to take deep inspiratory breath
IVC identified by junction with right atrium and/or presence of hepatic veins
M-mode not used, only B mode clips slowed to identify maximum and minimum diameter
Learn how to scan the IVC! 5 Minute Sono IVC
Recent article on what is the best way to measure and scan IVC (it was the way they did in this study)
Article on technical pitfalls of IVC scanning
Results
79 excluded due to various criteria
- 30 because IVF contraindicated
- 23 due to altered mental status
- 19 excluded from analysis because could not get images
N = 124
49.2% (61) were fluid responders
Differences between responders and nonresponders in demographics:
- more COPD in fluid responders
- more pulmonary HTN in nonresponders
Mean IVF of 4060mL prior to enrollment
Primary Results
Baseline cIVC had AUC 0.84
Optimal cIVC was 24.6%, rounded to 25%
sensitivity 87%
specificity 81%
+LR 4.56
-LR 0.16
Secondary Outcomes
Max or min baseline IVC did not detect fluid responsiveness
Change in cIVC with PLR did not detect fluid responsiveness
PLR + baseline cIVC did not help predict fluid responsiveness
Other Findings
PLR produced smaller change in cIVC in fluid responders compared to fluid nonresponders (opposite of what you would expect) → perhaps PLR did not stimulate the bolus or perhaps cIVC is just not a good measure of fluid responsiveness
False Positive (cIVC >25% but did not respond to IVF) – 4 patients with low EF or pulm HTN (authors states they have the capacitance for fluid but are not able to use it to augment stroke volume)
cIVC baseline:
Intrarater reliability: 0.92
Interrater reliability: 0.67
Supplemental Data:
- Baseline IVC inspiratory diameter – best cut off was ≥0.81 cm, yielding 71.4% sensitivity, and 16.4% specificity
- Baseline expiratory diameter was even worse.
Limitations
Is NICOM reliable gold standard?
Retrospectively derived cutoff of >25%
Although decent AUC, not sufficiently sensitive or specific to rule out or rule in
Exclusions! Many causes of shock were excluded, unclear how this would perform in undifferentiated shock population
Poor interrater reliability 0.67 – maybe this is hard to reproduce
Measured with clips and scrolling – not as easy as visual estimation
Patients got mean of over 4 liters of fluid prior to intervention – this shouldn’t affect the test itself but changes population to potentially have less fluid responders (affect external validity)
No patient centered outcomes – does it matter if cIVC predicts changes in cardiac index?
Take Home Points
1. In spontaneously breathing patients, an IVC collapsibility index >25% may have some predictability of volume responsiveness, but this data is not sturdy enough to rely on it definitely.
2. This study did not show improved predictability with dynamic measurements or passive leg raise which is contrary to prior literature.
Our score
