Comparison of Superior Vena Cava and Inferior Vena Cava Diameter Changes by Echocardiography in Predicting Fluid Responsiveness in Mechanically Ventilated Patients.

CONTEXT: Resuscitation of critically ill patients requires an accurate assessment of the patient's intravascular volume status. Passive leg raise cause auto transfusion of fluid to the thoracic cavity. AIMS: This study aims to assess and compare the efficacy of superior vena cava (SVC) and inferior vena cava (IVC) diameter changes in response to passive leg raise (PLR) in predicting fluid responsiveness in mechanically ventilated hemodynamically unstable critically ill patients.
METHODS: We enrolled 30 patients. Predictive indices were obtained by transesophageal and transthoracic echocardiography and were calculated as follows: (Dmax - Dmin)/Dmax for collapsibility index of SVC (cSVC) and (Dmax - Dmin)/Dmin for distensibility index of IVC (dIVC), where Dmax and Dmin are the maximal and minimal diameters of SVC and IVC. Measurements were performed at baseline and 1 min after PLR. Patients were divided into responders (increase in cardiac index (CI) ≥10%) and nonresponders (NR) (increase in CI <10% or no increase in CI).
RESULTS: Among those included, 24 (80%) patients were R and six were NR. There was significant rise in mean arterial pressure, decrease in heart rate, and decrease in mean cSVC from baseline to 1 min after PLR among responders. The best threshold values for discriminating R from NR was 35% for cSVC, with sensitivity and specificity of being 100%, and 25% for dIVC, with 54% sensitivity and 86.7% specificity. The areas under the receiver operating characteristic curves for cSVC and dIVC regarding the assessment of fluid responsiveness were 1.00 and 0.66, respectively.
CONCLUSIONS: cSVC had better sensitivity and specificity than dIVC in predicting fluid responsiveness. Copyright:

INTRODUCTION

The assessment of intra-vascular fluid status is very important in critically ill patients, particularly those suffering from septic shock.[1] Hemodynamic instability due to low cardiac output (CO), which in turn could be because of insufficient volume or septic myocardiopathy. Low CO in such patients may precipitate poor peripheral tissue perfusion and oxygenation.[2]

It has been shown in several studies that only about half of the critically ill patients who are also hemodynamically unstable are volume depleted.[3] This implies that the other half of the patients will not benefit, rather, may suffer potential or actual damage because of any further fluid administration, including any fluid challenge (7–15 ml.kg−1 body weight) as reported in previous studies. A patient is likely to benefit from fluid administration only when the heart has still some preload reserve (they are on the steep part of the Frank Starling curve).[456]

Most critically ill patients arrive at tertiary centers after being resuscitated in referring hospitals-either partially or fully. Of these patients, who are also hemodynamically unstable, ascribing the cause of such instability to volume depletion is difficult.[7]

Various static and dynamic parameters have been devised and studied with the sole aim of detecting volume depletion in critically ill hemodynamically unstable patients. Role of static parameters, for example, central venous pressure and pulmonary artery wedge pressure have been proved to be of very limited use, if not outright ineffective.[89] Dynamic parameters, using heart lung interaction have been suggested. They used respiration-induced changes in preload to determine the volume status. However, most of them (pulse pressure variation, stroke volume variation) suffer from limitations like using high tidal volumes (8–10 ml.kg−1) during measurements, need for sinus rhythm, and no spontaneous breathing effort.[1011]

Passive leg raising (PLR) induces a reversible autoshift of fluid from the lower limbs to the central compartment mimicking an exogenous fluid challenge, without the harmful effects of the former.[12]

Several measurements based on echocardiography like variations induced by respiration on the diameter of great vessels such as superior and inferior vena cava (IVC), to determine fluid responsiveness, had been proposed.[1314] We designed this study to evaluate and compare the efficacy of indices based on diameter changers of superior vena cava (SVC) and IVC by echocardiography in response to PLR in predicting fluid responsiveness in mechanically ventilated hemodynamically unstable patients of septic shock.

METHODS

This study was approved by the Institutional Review Board (IEC No-73/17) and written informed consent was obtained from the legal guardian of all subjects participating in the trial. The trial was registered prior to patient enrolment at Clinical Trial Registry of India (www.ctri.nic.in) (CTRI/2019/07/020480, date of registration: July 31, 2019). This was a monocentric, prospective, comparative study. We recruited hemodynamically unstable and critically ill patients, admitted to our intensive care unit (ICU). A total of 30 patients were included in the study. Of these 30 patients, 24 patients were responders and 6 were non-responders (NRs). The respiratory variations of the of collapsibility index SVC (cSVC) and distensibility index IVC (dIVC) were compared between hemodynamically unstable patients with or without vasopressor support after a passive leg raise in patients.

The patient was able to withdraw from the study at any time, without giving any reason and without any impact on treatment. A register was kept of all patients evaluated for inclusion and of patient who chose to withdraw from the study.

Inclusion criteria

All patients with suspected diagnosis of hypovolemic or septic shock on mechanical ventilator support admitted to ICU

Age between 18 and 60 years of age of either sex

The presence of at least one clinical sign of inadequate tissue perfusion defined as:

Systolic blood pressure (SBP) <90 mm of Hg or mean arterial blood pressure (MAP) < 65 mm of Hg

The need of vasopressor drugs

Urine output <0.5 ml.kg−1 h−1 for at least 2 h

Tachycardia (heart rate [HR] >100 min−1)

Serum Lactate (>2 mmol.L−1)

Central venous oxygen saturation (ScvO2) <70%.

Exclusion criteria

Patient's legal guardian's refusal to give consent

Patients with neurogenic, cardiogenic, or obstructive shock

Pregnancy

Amputation or severe limb ischemia

Contraindication to transesophageal echocardiography (TEE): Esophageal stricture, esophageal perforation, large diaphragmatic hernia, atlantoaxial disease, gastrointestinal bleeding, etc

Contraindication to passive leg raise (PLR) maneuver:-Elevated intracranial pressure, tamponade, acute aortic dissection, etc

Patients with intra-abdominal pressure more than 12 mm of Hg.

Sample size

The sample size was based on the previous study and calculated using the following formula:[15]

n = (Zα/2) 2 × p (1 − p)/d2 × Prevalence

Where n is the required sample size, P = Sensitivity, d = Precision, Zα/2 = Significance level

Taking 80% power, 5% significance level with 0.05 precision, the calculated sample size was 23.

n = (1.96 × 1.96) × 0.97 × 0.03/(0.05 × 0.05) × 0.52 = 23

Echocardiography

Transthoracic echocardiography (TTE) and TEE were performed on the patients by experienced intensivists trained in advanced critical care echocardiography. Because circulatory failure is frequently the reason for ICU admission, echocardiography was performed within the first 24 h of ICU admission. To allow safe introduction of esophageal probe, neuromuscular blocking agent was used along with suitable sedation in all patients

Collapsibility Index of superior vena cava (cSVC)

In mechanically ventilated patients, cSVC was calculated as (maximum diameter on expiration-minimum diameter on inspiration)/maximum diameter on expiration.[13]

cSVC or the collapsibility index of SVC (cSVC) = ([DmaxSVC − DminSVC]/DmaxSVC)

(Dmax and Dmin are the maximal and minimal diameters of SVC)

Distensibility Index of inferior vena cava (dIVC)

In mechanically ventilated patients, dIVC was calculated as, (maximum diameter on end inspiration-minimum diameter on end expiration)/minimum diameter on expiration).[14]

dIVC or the distensibility index of IVC = ([DmaxIVC − DminIVC]/DminIVC)

(Dmax and Dmin are the maximal and minimal diameters of IVC)

Velocity time integral

Velocity time integral (VTI) was measured by TTE at baseline and again at the end of the 1st min of PLR.

Cardiac output (CO)

CO= π × (LVOTd/2) 2 × VTI

LVOTd-left ventricular outflow tract diameter.[15]

Cardiac index

Cardiac index (CI)=CO/body surface area (BSA).

All the selected patients were mechanically ventilated using volume control mode with tidal volume 6–8 ml.kg−1, respiratory rate 12–15 min−1, PEEP up to 5 cm H2O, plateau pressure was kept below 30 cm H2O and intra-abdominal pressure ≤12 mm of Hg during the period of data collection. Ventilatory settings and dosages of ionotropic and vasopressor drugs were kept constant during entire study period. The patients remained sedated during the study using fentanyl 2 μg.kg−1 and midazolam (10–50 μg.kg−1) and paralyzed by using vecuronium (0.08–0.12 mg.kg−1) i. v. so that their spontaneous effort will be inhibited and to allow safe introduction of esophageal probe. A 22 G cannula was placed in radial artery and 7 Fr three lumen central venous catheter (CVC) was placed in the right internal jugular vein using Seldinger's technique after taking all aseptic precautions. Both radial artery and CVC were connected to transducers. Hemodynamic measurements were recorded with all the transducers positioned at the level of fourth intercostal space in the mid-axillary line. HR, SBP, diatolic blood pressure (DBP), MAP, the respiratory variation in SVC diameter measured by TEE and IVC diameter by TTE and change in CI measured by maximal Doppler velocity in the left ventricular outflow tract, were recorded at baseline (0 min) at 45° semi-recumbent position, 1 min after passive leg raise (PLR). PLR was done with elevated lower limbs at 45° while at the same time upper half of patient body in the supine. Three sets of measurements were recorded and averaged.

SVC diameter was measured by TEE in Bicaval view. SVC was identified as it enters the right atrium, M-mode applied and the M-mode cursor was placed to cross the SVC approximately 2 cm from the junction with the right atrium.

IVC diameter was measured by TTE after obtaining a subxyphoid view of the heart. With the ultrasound indicator on the probe directed toward the patient's left flank, the right atrium was identified. The ultrasound probe was then turned 90° counter clockwise. The IVC was identified as it enters the right atrium and M-mode applied. The M-mode cursor was placed across the IVC approximately 2 cm inferior to the junction with the right atrium.

The change in CI was measured by maximal Doppler velocity in the left ventricular outflow tract, aortic VTI measured with the help of TTE in apical 5-chamber view of the heart. Doppler mode was applied and cursor was placed just below the aortic valve.

The left ventricle outflow tract diameter (LVOTd) was measured with the help of TTE in para-sternal long axis view (PLAX view) in mid-systole, parallel to the aortic valve plane and within 0.5–1 cm of the valve orifice.

We used PLR to mimic fluid loading, because this maneuver has been shown to induce a reversible 500 ml of venous blood shift from the legs and the splanchnic reservoir to the thoracic compartment, which allows accurate identification of fluid responsiveness in ICU patients. PLR maneuver was done by elevating the lower limbs to 45° while the same time placing the patient in the supine from at 45° semi-recumbent position.[15]

At the end of the 1st min of PLR, patients were classified as fluid responders when the change in CI to passive leg raise recorded a rise in CI ≥10% whereas, those with an increase of <10% or no increase at all were classified as NR. The primary objective was to measure the cSVC, the distensibility index of IVC (dIVC) at baseline (0 min) and 1 min after PLR in a mechanically ventilated critically ill patients to classify patients as responders and NR and the secondary objectives were to compare and validate the accuracy and predictability of fluid responsiveness using cSVC, dIVC.

Statistical analysis

The results are presented in frequencies, percentages, and mean ± standard deviation. The Chi-square test was used to compare dichotomous/categorical variables between responders and NR. The unpaired t-test/Mann–Whitney U test was used to compare continuous variables between responders and NR at baseline and 1 min after PLR. The Wilcoxon rank sum test was used to compare change from baseline to 1 min. The receiving operating curve (ROC) analysis was carried out. The area under the curve (AUC) with its 95% confidence interval (CI) was calculated. The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) with their 95% CI were calculated. The P < 0.05 were considered statistically significant. All the analysis was carried out on SPSS 16.0 version (Chicago, Inc., USA).

RESULTS

Table 1 shows that there was no significant difference in demographic parameters such as age, gender (distribution of males), and anthropometric parameters (such as weight, height, body mass index, and body surface area) in responders and NR. Out of the 30 patients, vasopressor support was required to maintain hemodynamics in 17 (56.67%). Out of the 17 patients, 13 (76.5%) patients were responders and 4 (23.5%) were NR. There was no significant difference in the number of patients needing vasopressor support between the groups (P > 0.05) which could have confounded the findings.

Table 1

Comparison of demographic and anthropometric parameters and need for vasopressors in responders and nonresponders

Groups	Responders (n=24)	Nonresponders (n=6)	P
Age in years (mean±SD)	36.12±12.14	44.50±13.14	0.14
Gender (distribution of males) (%)	14 (77.8)	4 (22.2)	0.70
Weight (kg)	56.21±9.75	59.83±10.52	0.42
Height (cm)	162.50±3.48	165.17±1.94	0.08
BMI (kg/m2)	21.26±3.53	21.91±3.65	0.69
BSA (m2)	1.59±0.15	1.65±0.15	0.34
Vasopressor support (number of patients) (%)	13 (76.5)	4 (23.5)	0.58

*Statistically significant. BMI=Body mass index, BSA=Body surface area, SD=Standard deviation

Table 2 shows comparison of hemodynamic parameters (HR, mean arterial pressure), CI, collapsibility index SVC and distensibility index of IVC) at baseline and 1 min after PLR in responders and NR. cSVC was significantly different at baseline and 1 min after PLR between responders and NR.

Table 2

Comparison of hemodynamic parameters at baseline and one minute after passive leg raising in responders and nonresponders

Hemodynamic parameters	Baseline			One minute after PLR
	Responders (n=24)	Nonresponders (n=6)	P	Responders (n=24)	Nonresponders (n=6)	P
HR (beats/min)	100.79±13.32	102.22±20.08	0.83	94.25±11.74	97.17±16.32	0.61
MAP (mm of Hg)	70.48±6.28	73.20±10.84	0.42	75.06±7.88	78.04±9.00	0.43
CI (L/min/m2)	4.23±1.13	4.01±1.07	0.67	4.77±1.29	4.12±1.14	0.27
cSVC (%)	41.38±3.48	31.95±2.49	0.0001*	37.18±4.48	27.44±4.94	0.0001*
dIVC (%)	22.35±14.83	29.21±13.16	0.31	22.06±15.23	26.22±11.59	0.53

*Statistically significant (Unpaired t-test), †cSVC. dIVC=Distensibility index of inferior vena cava, PLR=Passive leg raising, HR=Heart rate, MAP=Mean arterial pressure, CI=Cardiac index, cSVC=Collapsibility index of superior vena cava

Table 3 shows that the mean change in hemodynamic parameters (systolic, diastolic and mean arterial pressure, HR, changes in superior and IVC diameters, left ventricular outlet diameter, VTI of the left ventricular outlet tract) from baseline to 1 min after PLR. There were significant changes in SBP, DBP, MAP, HR, ΔSVC, LVOTd, and LVOT-VTI from baseline to 1 min after PLR among responders.

Table 3

Comparison of mean change in hemodynamic parameters from baseline to 1 min among responders

Study parameters	Mean change (mean±SD)	P1
SBP (mm of Hg)	5.54±3.81	0.0001*
DBP (mm of Hg)	4.16±4.87	0.0001*
MAP (mm of Hg)	4.57±3.52	0.0001*
HR (beats/min)	6.54±3.34	0.0001*
cSVC (%)	4.19±3.23	0.0001*
dIVC (%)	0.28±5.68	0.80
LVOTd (mm)	0.05±0.04	0.0001*
LVOT-VTI (mm)	2.67±2.91	0.0001*
CI (L/min/m2)	0.53±0.18	0.0001*

Wilcoxon rank sum test, *Significant. cSVC=Collapsibility index of superior vena cava, dIVC=Distensibility index of inferior vena cava, LVOTd=Diameter of left ventricular outlet tract, LVOTVTI= Velocity time integral of left ventricular outlet tract, CI=Cardiac index, SBP=Systolic blood pressure, DBP=Diastolic blood pressure, MAP=Mean arterial pressure, HR=Heart rate, SD=Standard deviation

Table 4 shows the predictive values of cSVC for responders. cSVC >35% predicted responders correctly in 80% patients with sensitivity and specificity of being 100%.

Table 4

Predictive values of collapsibility index of superior vena cava for responders and nonresponders

	Responders, n (%)	Nonresponders, n (%)	Total, n (%)
cSVC cutoff
>35%	24 (80.0)	0 (0.0)	24 (80.0)
≤35%	0 (0.0)	6 (20.0)	6 (20.0)
Total	24 (80.0)	6 (20.0)	30 (100.0)
Predictive values, % (95% CI)
Sensitivity (%)	100.0 (100.0-100.0)
Specificity	100.0 (100.0-100.0)
PPV	100.0 (100.0-100.0)
NPV	100.0 (100.0-100.0)
AUC (95% CI)	1.00 (1.00-1.00)

Percentage from total number of cases. cSVC=Collapsibility index of superior vena cava, CI=Confidence interval, NPV=Negative predictive value, PPV=Positive predictive value, AUC=Area under the curve

Table 5 shows the predictive values of dIVC for responders. dIVC <25% predicted responders correctly in 43.3% patients with sensitivity and specificity of 66.7% (95% CI = 28.9–104.4) and 86.7% (95% CI = 69.5–103.9).

Table 5

Predictive values of distensibility index of inferior vena cava for responders and nonresponders

	Responders, n (%)	Nonresponders, n (%)	Total, n (%)
dIVC cutoff
<25%	13 (43.3)	2 (6.7)	15 (50.0)
≥25%	11 (36.7)	4 (13.3)	15 (50.0)
Total	24 (80.0)	6 (20.0)	30 (100.0)
Predictive values, % (95% CI)
Sensitivity	54.2 (34.2-74.1)
Specificity	66.7 (28.9-104.4)
PPV	86.7 (69.5-103.9)
NPV	26.7 (4.3-49.0)
AUC (95% CI)	0.66 (0.44-0.89)

Percentage from total number of cases. dIVC=Distensibility index of inferior vena cava, CI=Confidence interval, NPV=Negative predictive value, PPV=Positive predictive value, AUC=Area under the curve

Figure 1 depicts the ROC curve which shows the predictive values of cSVC for responders. cSVC >35% predicted responders correctly in 80% patients with sensitivity and specificity of 100%.

Figure 1

Receiving operating curve shows the predictive values of collapsibility index of superior vena cava for responders

Figure 2 depicts the ROC curve which shows the predictive values of dIVC for responders. dIVC <25% predicted responders correctly in 43.3% patients with sensitivity and specificity of 66.7% (95% CI = 28.9–104.4) and 86.7% (95% CI = 69.5–103.9) respectively.

Figure 2

Receiving operating curve shows the predictive values of distensibility index of inferior vena cava for responders

DISCUSSION

We report this study in which cSVC and dIVC were compared to predict fluid responsiveness in mechanically ventilated critically ill patients with sepsis. This is the only study till date which compared effectiveness of cSVC and dIVC without resorting to infusion of extraneous fluid (colloid or crystalloid).

There was significant difference in percentage increase in CI between responders and NR from baseline to 1-min after PLR [Table 3]. In previous studies also, it was reported that PLR increased aortic blood flow or CI by more than 10% to predict fluid responsiveness with a sensitivity of 97% and a specificity of 94% in mechanically ventilated patients considered for volume expansion.[1516]

There was significant rise in MAP from baseline to 1 min after PLR among responders [Table 3]. Several studies reported that MAP could be used as an indicator of fluid responsiveness in patients with septic shock and on mechanical ventilation.[17181920] In contrast to our findings, a previous study reported that fluid challenge did not improve MAP in about one-third of hypotensive critically ill mechanically ventilated patients. They justified that in preload-dependent patients, CO can initially increase by a large extent despite minimal arterial pressure increase because of decrease in SVR.[21]

There was significant decrease in the mean HR from baseline to 1 min after PLR among responders [Table 3]. This finding was similar to those of previous studies.[2022]

There was significant difference in cSVC between responders and NR at baseline and 1-min after PLR [Table 2] and also there was a significant decrease in the mean cSVC from baseline to 1 min after PLR among responders [Table 3]. In our study, the cSVC >35% predicted responders correctly in 80% patients with sensitivity and specificity of 100% each, and PPV and NPV of 100% each [Table 4]. These findings were similar to the one in which volume status in mechanically ventilated septic patients were gauged by cSVC after volume expansion by HES. The threshold cSVC of 36%, allowed discrimination between responders and NR, with a sensitivity of 90% and a specificity of 100%.[13] In another study, it was reported that the best cutoff value of cSVC to predict fluid responsiveness was 29% with a sensitivity of 54% and a specificity of 94%.[23] Although findings are similar, our study was different from these previous studies because, in them hydroxyethyl starch solution (HES) was infused for volume expansion, included septic shock patients with associated acute lung injury (ALI) and used higher tidal volumes (8-10 ml.kg−1). Another recent study found that cSVC ≥21% predicted responders with a sensitivity of 61% and a specificity of 84%.[24] They used PLR and included patients of circulatory failure of various causes and mechanically ventilated with 8–10 ml.kg−1.

In our study, we found that there was no significant difference in dIVC neither of responders and NR at baseline and 1-min after PLR [Table 2] nor in the mean change in dIVC from baseline to 1 min after PLR among responders and NR [Table 3]. In our study, the dIVC >25% predicted responders correctly in 43.3% patients with modest sensitivity and specificity of 54.2% and 66.7% respectively and a PPV of 86.7% and a NPV of 26.7% [Table 5]. One study involving patients with acute circulatory failure related to sepsis reported that using a threshold dIVC of 18%, responders and NR were discriminated with 90% sensitivity and 90% specificity.[14] A strong relation (r = 0.9) was observed between dIVC at baseline and the CI increase following blood volume expansion.[14] Another study involving mechanically ventilated patients with septic shock, reported that volume loading induced a significant increase in CO and a decrease in dIVC and dIVC cut-off value of 12% allowed identification of responders with PPV and NPV of 93% and 92%, respectively.[25] In one study, it was found that the best cut off value of dIVC was 21% with a sensitivity of 38% and a specificity of 61% (23). The previous studies were different as they utilized volume loading with HES, whereas, we employed volume shift due to PLR.[142325]

Several previous studies also reported poor sensitivity and specificity of dIVC in predicting responders to fluid challenge. In another, it was reported that dIVC ≥ 8% predicted responders with a sensitivity of 55% and a specificity of 70%.[24] Our study results were different from this study because they included all patients with acute circulatory failure of any cause and used higher tidal volume (>8 ml.kg−1) in a subset of patients.

We found that the AUC operating characteristic curves for cSVC and dIVC was 1.0 and 0.66, respectively. This showed that cSVC had better accuracy compared to dIVC regarding the assessment of fluid responsiveness and there was no significant correlation between cSVC and dIVC among responders and NR at baseline [Tables 4 and 5]. Literature reveals only two studies which compared cSVC and dIVC to predict fluid responsiveness in mechanically ventilated critically ill patients. In the first study, it was reported that the AUC for cSVC (0.74) was more than that of dIVC (0.43).[23] No significant correlation between cSVC and dIVC was found. In the other, it was found that the AUC of cSVC (0.755) was significantly greater than that of dIVC (0.635).[24] Although the findings of our study and these[2324] concur that cSVC was better than dIVC as a predictor of fluid responsiveness in mechanically ventilated hemodynamically unstable critically ill patients, the study design (method of volume expansion-exogenous or autologous by PLR) and selection of patients were different in these studies.

Our study suffers from the following limitations:

The sample size of the study was small (30 patients)

We recruited only a small subset of critically ill patients, namely, septic shock, and we excluded patients of ALI and raised IAP.

Hence, we suggest that further studies with larger sample size and including hemodynamically unstable patients of various etiologies may be conducted.

CONCLUSIONS

cSVC should systematically be measured during routine echocardiography in septic shock as it gives an accurate index of fluid responsiveness.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.