Sage Journals: Discover world-class research

Abstract

Objectives:

To systematically review the diagnostic utility of the respiratory variation of the inferior vena cava diameter measured using ultrasonography for predicting fluid responsiveness in adult patients and compare the three commonly used equations, inferior vena cava distensibility, inferior vena cava collapsibility and inferior vena cava variability.

Methods:

We searched PubMed, Scopus, Web of Science and Cochrane library, and included studies investigating the diagnostic accuracy of the respiratory variation of the inferior vena cava measured using ultrasonography compared to a reference standard for measuring cardiac output after a fluid challenge for fluid responsiveness, and stratified participants as fluid responsive or not. We included studies conducted in the emergency department or intensive care unit. We excluded studies on paediatric, prehospital, cancer, pregnant, dialysis patients or healthy volunteers.

Results:

We retrieved 270 records and excluded 171 because of irrelevance, patient population or publication type. We screened the abstracts of 99 studies and then the full texts of 42 studies. Overall, 21 studies with 1321 patients were included, of whom 689 (52%) were fluid responsive. The mean threshold value for positive inferior vena cava distensibility, inferior vena cava collapsibility and inferior vena cava variability was 17%, 35% and 12%, respectively. The heterogeneity between studies was high. Bivariate diagnostic random-effects meta-analysis was used to calculate the summary receiver operating characteristics curves. The overall accuracy, sensitivity and specificity of respiratory variation of the inferior vena cava diameter were 0.85, 0.72 and 0.81, respectively. The accuracy of inferior vena cava distensibility and inferior vena cava collapsibility was similar. The diagnostic utility of respiratory variation of the inferior vena cava diameter was lower but not statistically significant in mechanically ventilated patients compared with spontaneous breathing for predicting fluid responsiveness.

Conclusion:

The respiratory variation of the inferior vena cava diameter has moderate diagnostic utility for predicting fluid responsiveness independent of the equation used.

Keywords

Respiratory variation inferior vena cava fluid responsiveness meta-analysis

Introduction

In hemodynamically unstable critical care patients, intravenous fluids are administered to increase cardiac output (CO), tissue perfusion and systemic blood pressure.^1,2 However, there seems to be a threshold where more fluid administration does not increase the stroke volume (SV), and harmful effects, such as pulmonary or peripheral oedema, are observed.^1,2 This concept is called fluid responsiveness (FR), and static and dynamic measures were introduced to estimate whether a patient is fluid responsive. Dynamic measures represent changes in CO, SV, or similar parameters after a manoeuvre, such as passive leg raise (PLR) or positive pressure breaths. In PLR, the lower extremities were elevated to transfer pooled blood to the central compartment to increase preload, where an increase in any hemodynamic measure suggests FR.

Right atrial and central venous pressures tend to change with respiration, and those changes also change the diameter of the inferior vena cava (IVC). The respiratory variation of IVC can be measured using ultrasonography (US) at the bedside, which may help estimate the patient’s FR. Previous studies assessing the diagnostic accuracy of the respiratory variation of the IVC used different static or dynamic measures as the reference standard, used different thresholds, were conducted on remarkably different patient populations, and reported the results of three equations with the same numerator that is normalised to slightly different denominators. Therefore, the meta-analysis of those studies showed considerable heterogeneity with mixed results. Thus, we aimed to systematically review the diagnostic utility of the respiratory variation of the IVC measured using USG for FR in adult patients and compare the utility of inferior vena cava distensibility (IVCd), inferior vena cava collapsibility (IVCc) and inferior vena cava variability (IVCv) equations.

Methods

This study complied with the recent update of the preferred reporting items for systematic review and meta-analysis of diagnostic test accuracy studies (PRISMA-DTA) and PRISMA-DTA for abstracts.³ The study and search protocols were not registered or published. Written informed consent was not necessary because patient data were excluded.

Criteria for study selection

Types of studies

We included prospective observational diagnostic accuracy studies with a reference standard for measuring CO after a fluid challenge and those stratifying participants into FR and non-FR. We reported the number of true positive (TP), false positive (FP), false negative (FN) and true negative (TN) patients, sensitivity or specificity of IVCd ((Dmax − Dmin)/Dmin), IVCc ((Dmax − Dmin)/Dmax) or IVCv ((Dmax − Dmin)/(Dmax + Dmin)/2) indexes measured using USG for predicting FR, where Dmax is the maximal and Dmin is the minimal IVC diameter. We considered studies published in peer-reviewed journals in English and excluded studies with a case–control design.

Participants

We included studies conducted in the setting of an emergency department (ED) or intensive care unit (ICU), involving adult (age ⩾18 years) participants of any sex. Moreover, studies involving both mechanically ventilated (MV) and spontaneously ventilating patients were included. We excluded studies including the paediatric population, prehospital setting, patients with cancer, pregnant patients, patients on dialysis or healthy volunteers.

Index tests

We included studies that measured the IVC diameter in M-mode from the subxiphoid view, from the longitudinal axis. The maximum allowed time between IVC US and the fluid challenge was 30 min. We excluded studies measuring IVC from mid-axillary views or with a lateral approach and calculated area or cross-sectional diameters. We also excluded studies with missing equations used for calculating the IVCv index. Blinding of the sonographers to the SV, CO or velocity–time integral (VTI) change was not always feasible and practical. Therefore, we graded this point during the quality assessment.

Target condition and reference standards

We included studies that reported FR with the change in SV, stroke index (SI), CO or cardiac index after a fluid challenge or PLR, with any of the following well-established techniques: transthoracic echocardiography (TTE), transpulmonary thermodilution (TPTD), arterial pulse waveform analysis (APWA), bioreactance (BR), pulse contour analysis (PCA) or transoesophageal echocardiography (TEE). We expected considerable heterogeneity in defining FR a priori.

Search methods for identification of studies

Electronic search

We searched The Cochrane Database for Systematic Reviews, the National Institute for Health and Care Excellence (NICE) and Epistemonikos for existing reviews based on methods outlined elsewhere.⁴ We searched the MEDLINE, Embase and Web of Science databases from their start to June 2020 using the Medical Subject Headings (MeSH) and natural language words for IVC, FR/challenge and ultrasound.

The search query used for PubMed was as follows: ((‘vena cava, inferior’[MeSH Major Topic] OR ‘inferior vena cava’[Title/Abstract]) AND (((‘fluid responsiveness’[Title/Abstract] OR ‘volume expansion’[Title/Abstract]) OR ‘fluid challenge’[Title/Abstract]) OR ‘preload assessment’[Title/Abstract])) AND (‘ultrasonography’[MeSH Major Topic] OR (‘ultraso*’[Title/Abstract] OR ‘sonogr*’[Title/Abstract])).

Searching other resources

We reviewed the reference sections of the relevant original articles and reviews for footnote chasing. We excluded reviews, editorials, case reports, letters to the editors, correspondences, conference abstracts, non-English studies and non-human studies.

Data collection and analysis

We collected the data as described in The Cochrane Handbook for Systematic Reviews of Diagnostic Test Accuracy guidelines.⁵

Study eligibility and quality assessment

Initial searches at the above databases were exported to a reference manager file and imported to the online systematic review search app Rayyan QCRI.⁶ The duplicates were checked and automatically removed. Subsequently, we screened the title and abstracts of the batch for publication type, population and relevance (E.U.A. and H.A.). After the screening, we retrieved the papers’ full text to evaluate their eligibility (E.U.A. and H.A.). We resolved disagreements on study eligibility by discussion. We exported our final list of studies to RevMan 5. Figure 1 shows our process flow.

Figure 1.

Flow chart of the systematic review process.

Data extraction and management

We collected data on four domains into an electronic spreadsheet: (1) study characteristics: title, authors, country, publication year, design, language, setting and number of centres; (2) patient population: total number of patients, and number of patients in the study and control groups, the age range of the study population, ventilation status (spontaneous or mechanical ventilation), tidal volume and positive end-expiratory pressure (PEEP) if MV was used; (3) index test: IVC formula, calculated or pre-defined cut-off value; (4) reference test: definition of FR, the pre-defined threshold for FR, the metric and device used, PLR or fluid challenge, volume and composition of the fluid; (5) outcomes: area under the curve (AUC) of receiver operating characteristics (ROC) for FR at the defined threshold value of the index test with TP, TN, FP, TN values, sensitivity and specificity.

Assessment of methodological quality

We evaluated the design and reporting quality of all included studies using the Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2).^7,8 We recorded our evaluations in an excel template that we downloaded from the RobVis tool website.⁹ Subsequently, we used RevMan 5 to create the risk of bias and applicability concerns graphs (Figure 2).

Figure 2.

Risk of bias and applicability concerns: (a) graph and (b) summary: review authors’ judgements about each domain presented as percentages across included studies.

Statistical analysis and data synthesis

We calculated the summary statistics for diagnostic accuracy from TP, TN, FP and TN values. We created the probability plot, forest plots, Deek’s funnel plot and Fagan nomogram with the Stata statistical software’s midas function. We performed a bivariate diagnostic random-effects meta-analysis described by Reitsma et al.¹⁰ to calculate the summary receiver operating characteristic (SROC) curves and plotted them with their associated confidence regions and summary points. During this review, we used RevMan 5 (computer program) (Version 5.4, Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration (2014)), RStudio (R Core Team, Vienna, Austria; https://www.R-project.org/) and Stata (StataCorp LLC, TX, USA) for the statistical analyses. Statistical significance was considered when two-tailed p < 0.05.

Investigations of heterogeneity

We expected significant heterogeneity due to differences in study populations, settings, variation in IVC measurement method, threshold values, and variation in the reference standard, and its threshold value for defining FR. We evaluated the heterogeneity using Cochrane’s Q test, bivariate version of the Higgins’ I², reported with each forest plot, and considered significant when Q < 0.1 and I² > 50%.

Sensitivity analysis

We created a probability-modifying plot, a graphical sensitivity analysis of predictive value across a prevalence continuum defining low- to high-risk populations (Figure 3(a)).

Figure 3.

(a) The probability-modifying plot. (b) Deek’s funnel plot for the assessment of publication bias.

Assessment of reporting bias

We used Deek’s version of the funnel plot to evaluate reporting bias. We graphed the regression of diagnostic log odds ratio against 1/sqrt (effective sample size), weighting by effective sample size, where p < 0.10 for the slope coefficient indicating significant asymmetry and no significant asymmetry (Figure 3(b)).

Results

Search results

We retrieved 270 records from the databases’ initial search and reference chasing and excluded 171 records due to irrelevance, wrong patient population or publication type. Figure 1 shows the study selection process. We screened the abstracts of 99 studies and full texts of 42 studies for eligibility. After excluding 78 more studies, 21 studies fulfilled all inclusion criteria and were included in the analysis.

The methodological quality of the included studies

Tables 1 –3 show the main characteristics of the included studies. We grouped studies according to the equation used. Of the 21 studies, 11, 8 and 3 studies reported IVCd, IVCc and IVCv index, respectively. One study was conducted in the operating room (OR), one in the ED and the rest in the ICU. Only one study included patients aged >14 years; the rest included adults only. Moreover, 8 and 13 studies included sedated patients on MV and those with spontaneous breathing. The calculated threshold values of the IVC indexes for FR varied among studies. The site of IVC diameter measurement, fluid challenge volume and content was also variable among studies. The reference standard tests for FR were SV, stroke volume index (SVI), stroke volume variation (SVV), CO, cardiac index (CI) or VTI. In 13 studies, the reference standard was measured using TTE, and FR was defined as an increase of >10%–15% following fluid challenge or PLR. In the remaining studies, the reference standard tests were measured using PCA, APWA, BR or TPTD with variable thresholds. A threshold of 15% and 10% for FR was accepted in 11 and 10 studies, respectively.

Table 1.

Characteristics of studies included in the meta-analyses: IVC distensibility (n = 11).

Author	Year	Country	Design	Patient group	MV	Setting	TV, mL/kg	PEEP, cmH₂O	Fluid challenge and volume	FR measurement and device	FR threshold	Index test measurement method	Comment
Barbier et al.¹¹	2004	France	Prospective	Acute circulatory failure related to severe sepsis	Yes	ICU	8.5	4	7 mL/kg, 4% modified fluid gelatin 30 min	CI, TTE LVOT	15%
Charbonneau et al.¹²	2014	France	Prospective	Acute circulatory failure related to septic shock	Yes, sedated	ICU	8–10	Not available	Not available	CI, TEE LVOT	⩾15%		Patients with arrhythmia excluded
de Oliveira et al.¹³	2016	Brazil	Prospective	Post-op surgical patients with clinical requirement for a fluid challenge according to the attending physician	Yes	ICU	8	Not set a priori	500 cc crystalloid 15 min	VTI, TTE LVOT	⩾15%		Blinding unclear
Lu et al.¹⁴	2017	P.R.C.	Prospective	Septic shock	Yes	ICU	8–10	5–12	200 cc NS 10 min	CI, TPTD (PiCCO)	⩾10%		Blinding unclear
Machare-Delgado et al.¹⁵	2011	USA	Prospective	Vasopressor-dependent patients requiring fluids due to hypotension or poor organ perfusion	Yes	ICU	8	Not set a priori	500 cc NS 10 min	SVI, TTE LVOT VTI	⩾10%		Pre-specified index threshold
Moretti and Pizzi ¹⁶	2010	Italy	Prospective, consecutive	Subarachnoid haemorrhage and low cardiac index	Yes	OR	8	No PEEP	7 mL/kg, 6% HES 30 min	CI, TPTD	15%		Patient with SAHNot shock but low patients with low CI
Sobczyk et al.¹⁷	2016	Poland	Prospective, consecutive	Post-operative fluid management after elective coronary artery bypass grafting	Yes	ICU	Unavailable	Unavailable	250 cc NS 10 min	CO, TTE LVOT	⩾15%		Pre-specified index threshold
Theerawit et al.¹⁸	2016	Thailand	Prospective	Severe sepsis or septic shock requiring intravenous fluid challenge	Yes	ICU	>8	8–10	1000 cc NS in 1 h, 500 cc 6% HES in 30 min or 5% alb in 30 min	SVV, APWA (Vigileo + FloTrac)	15%		Sampling method unclear
Vignon et al.¹⁹	2017	France	Prospective, consecutive	A subgroup of patients with septic shock	Yes	ICU	Variable	Variable	PLR	VTI, TTE LVOT	10%
Yao et al.²⁰	2019	P.R.C.	Prospective, unknown sampling method	Mixed, 18% with shock	Yes	ICU	Not set a priori	Not set a priori	PLR	CO, APWA (CNAP)	>10%	IVC cross-sectional area and diameter, not longitudinal axis	Sampling method unclearBlinding unclearMost patients did not require fluid replacement
Zhang et al.²¹	2019	P.R.C.	Prospective	Mixed patients with hypotension, tachycardia, oliguria or hyperlactatemia. Post-cardiac surgery patients excluded	Yes	ICU	Not set a priori	Not set a priori	PLR	VTI, TTE LVOT	10%		Blinding unclearExclusion for inadequate imaging (83/268, 31%)

IVC distensibility: (Dmax − Dmin)/Dmin.

APWA: arterial pulse waveform analysis; BR: bioreactance; CI: cardiac index; CNAP: continuous noninvasive arterial pressure; CO: cardiac output; FR: fluid responsiveness; HES: hydroxyethyl starch; ICU: intensive care unit; IVC: inferior vena cava; LVOT: left ventricle outflow tract; MV: mechanical ventilation; OR: operating room; PEEP: post-expiratory end-pressure; PLR: passive leg raise; SVI: stroke volume index; SVV: stroke volume variation; SAH: subarachnoid haemorrhage; TEE: transoesophageal echocardiography; TPTD: transpulmonary thermodilution; TTE: transthoracic echocardiogram; TV: tidal volume; VTI: velocity–time integral.

Table 2.

Characteristics of studies included in the meta-analyses: IVC collapsibility (n = 8).

Author	Year	Country	Design	Patient group	MV	Setting	Fluid challenge and volume	FR measurement and device	FR threshold	Comment
Airapetian et al.²²	2015	France	Prospective, consecutive	Clinical signs of acute circulatory failure, dehydration or oligoanuria	No	ICU	500 cc NS 15 min	CO, TTE LVOT VTI	⩾10%	Blinding unclear
Bortolotti et al.²³	2018	Italy	Prospective	Infection-related acute circulatory failure and irregular cardiac rhythm	No	ICU	500 cc 4% gelatin	SVI, TTE LVOT VTI	⩾10%	Patients with irregular cardiac rhythm were included onlyIVC-CI measured during spontaneous (included) and standardised breaths30/85 (35.3%) patients excluded for unmeasurable parameters
Corl et al.²⁴	2017	USA	Prospective, convenience sample	Critically ill patients undergoing resuscitation	No	ICU/ED	500 cc NS	CI, BR	⩾10%	High rate of patients declined study inclusionConvenience samplingTraumatic, cardiogenic, obstructive or neurogenic shock excluded
Corl et al.²⁵	2019	USA	Prospective	Signs of acute circulatory failure at any time within 72 h of presentation to the ED	No	ICU	500 cc NS	CI, BR	⩾10%	Novice vs expert (included) sonographersHigh rate of patients declined study inclusion
Lanspa et al.²⁶	2013	USA	Prospective	Septic shock	No	ICU	10 mL/kg crystalloid in 20 min	CI, TTE LVOT VTI	⩾15%	Sampling method unclearSmall sample sizeBlinding unclear
McGregor et al.²⁷	2020	UK	Prospective		No	ED	250–500 cc NS 15 min	SV, TTE LVOT VTI	⩾10%	Novice sonographers
Muller et al.²⁸	2012	France	Prospective	Acute circulatory failure	No	ICU	500 cc 6% HES 15 min	VTI, TTE LVOT	⩾15%	No blindingReported values do not match
Preau et al.²⁹	2017	France	Prospective, consecutive	Sepsis-induced acute circulatory failure and considered for volume expansion	No	ICU	500 cc 4% gelatin 30 min	SVI, TTE LVOT VTI	⩾10%	IVC-CI at spontaneous (included) and standardised breaths22/112 (19.6%) patients excluded for unmeasurable parameters

IVC collapsibility: (Dmax − Dmin)/Dmax.

BR: bioreactance; CI: cardiac index; ED: emergency department; FR: fluid responsiveness; ICU: intensive care unit; IVC: inferior vena cava; MV: mechanical ventilation; SV: stroke volume; SVI: stroke volume index; TTE: transthoracic echocardiogram; VTI: velocity–time integral; HES: hydroxyethyl starch.

Table 3.

Characteristics of studies included in the meta-analyses: IVC variability (n = 3).

Author	Year	Country	Design	Patient group	MV	Setting	TV, mL/kg	PEEP, cmH₂O	Fluid challenge and volume	FR measurement	FR threshold	Comments
Feissel et al.³⁰	2004	USA	Prospective	Patients with septic shock with clinical signs of acute circulatory failure or/and biological signs of organ dysfunction	Yes	ICU	8–10	Not available	8 mL/kg 6% HES 20 min	CO, TTE LVOT VTI	⩾15%	Sampling method unclear
Ma et al.³¹	2018	P.R.C.	Prospective	Patients with circulatory instability required rapid fluid challenge	Yes, sedated, no NMB	ICU	8	5	500 cc HES 30 min	SVV, APWA (Vigileo + FloTrac)	⩾15%	Patients with arrhythmia excluded
Theerawit et al.¹⁸	2016	Thailand	Prospective	Severe sepsis or septic shock requiring intravenous fluid challenge	Yes	ICU	>8	8–10	1000 cc NS in 1 h, 500 cc 6% HES in 30 min or 5% alb in 30 min	CO, PCA	15%	Sampling method unclear

IVC variability: (Dmax − Dmin)/((Dmax − Dmin)/2).

APWA: arterial pulse waveform analysis; CI: cardiac index; CO: cardiac output; FR: fluid responsiveness; HES: hydroxyethyl starch; ICU: intensive care unit; IVC: inferior vena cava; MV: mechanical ventilation; NMB: neuromuscular blocker; TV: tidal volume; PCA: pulse contour analysis; PEEP: post-expiratory end-pressure; SV: stroke volume; TTE: transthoracic echocardiogram.

The quality of the included studies was assessed using QUADAS-2 (Figure 2). All studies met the inclusion criteria for the index test, reference standard and patient population. The duration between the index and reference tests was within the acceptable range, and both the index and reference tests were performed in all included patients. Information on the consecutive or random sampling of the participants was consistently missing in almost all studies. Information regarding the blinding of the performers and interpreters of the index and reference tests were either missing or unavailable in most studies. Inter- and intra-observer variability in the index test and the reference test was also poorly reported.

Findings

Overall, 21 studies with 1321 patients were included, of whom 689 (52%) were fluid responsive (Tables 4 –6). The mean threshold values for a positive index test were 17%, 35% and 12% for IVCd, IVCc and IVCv, respectively.

Table 4.

Summary of findings: IVC distensibility (n = 11).

Author	Year	Country	Total sample size, N	Fluid responsive, n (%)	IVC threshold	AUC (95% CI)	Sensitivity (95% CI)	Specificity (95% CI)	TP	FP	FN	TN
Barbier et al.¹¹	2004	France	20	10 (50%)	18%	0.91 (0.84–0.98)	0.90 (0.55–1.00)	0.90 (0.55–1.00)	9	1	1	9
Charbonneau et al.¹²	2014	France	44	26 (59%)	21%	0.43 (0.25–0.61)	0.38 (0.20–0.59)	0.61 (0.36–0.83)	10	7	16	11
de Oliveira et al.¹³	2016	Brazil	20	9 (45%)	16%	0.84 (0.63–1.00)	0.67 (0.30–0.93)	1.00 (0.72–1.00)	6	0	3	11
Lu et al.¹⁴	2017	P.R.C.	49	27	20.50%	0.81 (0.67–0.94)	0.67 (0.46–0.83)	0.77 (0.55–0.92)	18	5	9	17
Machare-Delgado et al.¹⁵	2011	USA	25	8 (32%)	⩾12% (PS)	0.81 (0.64–0.99)	1.00 (0.63–1.00)	0.53 (0.28–0.77)	8	8	0	9
Moretti and Pizzi²²	2010	Italy	29	17 (59%)	>16%	0.90 (0.73–0.98)	0.71 (0.44–0.90)	1.00 (0.74–1.00)	12	0	5	12
Sobczyk et al.¹⁷	2016	Poland	35	24 (68.6%)	18% (PS)	0.74	0.82 (0.63–0.95)	0.73 (0.39–0.94)	20	3	4	8
Theerawit et al.¹⁸	2016	Thailand	29	16 (55.2%)	10.7%	0.69 (0.48–0.90)	0.75 (0.48–0.93)	0.77 (0.46–0.95)	12	3	4	10
Vignon et al.¹⁹	2017	France	236	128 (54.2%)	8%	0.63	0.53 (0.44–0.62)	0.74 (0.65–0.82)	68	28	60	80
Yao et al.²⁰	2019	P.R.C.	67	37 (55.2%)	25.6%	0.70 (0.58–0.83)	0.46 (0.29–0.63)	0.90 (0.73–0.98)	17	3	20	27
Zhang et al.²¹	2019	P.R.C.	129	62 (48%)	16.5%	0.82 (0.74–0.89)	0.79 (0.67–0.88)	0.72 (0.60–0.82)	49	19	13	48
Total			683	364					229	77	135	242

AUC: area under the curve; TP: true positive; FP: false positive; FN: false negative; TN: true negative; CI: cardiac index; IVC: inferior vena cava; NA: not available, PS: pre-specified.

Table 5.

Summary of findings: IVC collapsibility (n = 8).

Author	Year	Country	Total sample size, N	Fluid responsive, n (%)	IVC threshold	AUC (95% CI)	Sensitivity (95% CI)	Specificity (95% CI)	TP	FP	FN	TN
Airapetian et al.²²	2015	France	59	29 (49%)	42%	0.62 (0.49–0.74)	0.31 (0.15–0.51)	0.97 (0.83–1.00)	9	1	20	29
Bortolotti²⁹	2018	Italy	55	29 (53%)	39%	0.82 (0.70–0.93)	0.66 (0.46–0.82)	0.85 (0.65–0.96)	19	5	10	26
Corl et al.²⁴	2017	USA	124	61 (49.2)	25%	0.84 (0.76–0.91)	0.87 (0.76–0.94)	0.81 (0.69–0.90)	53	12	8	51
Corl et al.²⁵	2019	USA	85	44 (52%)	25% (PS)	0.82 (0.74–0.88)	0.86 (0.73–0.94)	0.78 (0.63–0.88)	38	9	6	32
Lanspa et al.²⁶	2013	USA	14	5 (36%)	NA	0.833 (0.58–1.00)	1.00 (0.48–1.00)	0.67 (0.30–0.93)	5	3	0	6
McGregor et al.²⁷	2020	UK	33	20 (60.6%)	40%	0.464 (0.264–.675)	0.47 (0.24–0.71)	0.64 (0.31–0.89)	9	4	10	7
Muller et al.²⁸	2012	France	40	20 (50%)	40%	0.77 (0.60–0.88)	0.70 (0.46–0.88)	0.80 (0.56–0.94)	14	4	6	16
Preau et al.²⁹	2017	France	90	50 (56%)	⩾31%	0.82 (0.73–0.91)	0.76 (0.62–0.87)	0.88 (0.73–0.96)	38	5	12	35
Total			500	258					185	43	72	202

AUC: area under the curve; TP: true positive; FP: false positive; FN: false negative; TN: true negative; CI: cardiac index; IVC: inferior vena cava; NA: not available, PS: pre-specified.

Table 6.

Summary of findings: IVC variability (n = 3).

Author	Year	Country	Total sample size, N	Fluid responsive, n (%)	IVC threshold	AUC (95% CI)	Sensitivity (95% CI)	Specificity (95% CI)	TP	FP	FN	TN
Feissel et al.³⁰	2004	USA	39	16 (41%)	12%	NA	0.81 (0.54–0.96)	0.96 (0.78–1.00)	13	1	3	22
Ma et al.³¹	2018	P.R.C.	70	35 (50%)	13.4%	0.83 (0.72–0.91)	0.86 (0.70–0.95)	0.86 (0.70–0.95)	30	5	5	30
Theerawit²⁴	2016	Thailand	29	16 (55.2%)	10.2%	0.688 (0.480–0.895)	0.75 (0.48–0.93)	0.77 (0.46–0.95)	12	3	4	10
Total			138	67					55	9	12	62

AUC: area under the curve; TP: true positive; FP: false positive; FN: false negative; TN: true negative; CI: cardiac index; IVC: inferior vena cava; NA: not available, PS: pre-specified.

The overall diagnostic accuracy (AUC of the SROC curve) of the respiratory variability of the IVC diameter was 0.85 (95% confidence interval (CI) = 0.81–0.87) with sensitivity and specificity of 0.72 (95% CI = 0.64–0.79) and 0.81 (95% CI = 0.76–0.86), respectively (Figure 4(a)). The diagnostic accuracies of IVCc and IVCd were 0.86 (95% CI = 0.83–0.89) and 0.81 (95% CI = 0.77–0.84), with sensitivities of 0.73 (95% CI = 0.56–0.85) and 0.69 (95% CI = 0.57–0.79), and specificities of 0.83 (95% CI = 0.77–0.88) and 0.78 (95% CI = 0.68–0.85) (Figure 4(b) and (c)). The 95% confidence contours of the SROC curves of the distensibility and collapsibility indexes were overlapped, suggesting similar diagnostic accuracies. Because only three studies reported IVCv, creating an SROC curve was not possible.

Figure 4.

Summary ROC plot of tests with summary point and confidence regions: (a) all studies, (b) studies of IVC collapsibility and (c) studies of IVC distensibility. Circles represent each study included in the meta-analysis.

The forest plot of the pooled sensitivity and specificities (Figure 5) and positive and negative diagnostic likelihood ratios (DLRs) (Figure 6) were presented according to the equations. The pooled positive and negative DLRs of the IVC diameter were 3.86 (95% CI = 2.97–5.01) and 0.34 (95% CI = 0.26–0.45), respectively, suggesting moderate diagnostic utility for predicting the presence of FR. The heterogeneity between studies was low for positive likelihood ratios but high for negative likelihood ratios.

Figure 5.

Forest plot of the pooled sensitivity and specificities: (a) all studies, (b) studies reporting IVC collapsibility and (c) studies reporting IVC distensibility.

Figure 6.

Forest plot of the pooled positive and negative diagnostic likelihood ratios: (a) all studies, (b) studies reporting IVC collapsibility and (c) studies reporting IVC distensibility.

Meta-regression analysis showed no significant difference between the sensitivity and specificities of subgroups according to the setting (ED, ICU) or ventilation (spontaneous, MV). The clinical utility of the respiratory variation of the IVC diameter was graphed for the pre-test probability of 25% in the Fagan nomogram in Figure 7. In a patient with a pre-test probability of 25%, the post-test probability of FR would be 56% and 10% in the presence or absence of IVCv, respectively.

Figure 7.

Fagan nomogram.

Summary of main results

This meta-analysis including 21 studies with 1321 patients showed that the respiratory variation of the IVC diameter had moderate diagnostic utility for predicting FR, regardless of the equation used. We did not find any significant difference in the subgroups because all confidence regions were overlapped, and only three studies were included in the IVCv domain. For IVCc, positive DLR was homogeneous among studies included. The heterogeneity of the included studies was high.

Discussion

One of the essential tasks of providers is predicting the FR of critical care patients in the ED or ICU. Therefore, those indexes still have great value in practice. Studies on the use of respiratory variability of the IVC diameter for predicting FR concluded discordant results despite the increase in their methodological quality and sample size in recent years.³² The respiratory variation of the IVC diameter had a moderate diagnostic utility to predict FR. There were paucity of data regarding the selection of the optimal equation to calculate the variability of the IVC diameter. The diagnostic utilities of different equations did not differ significantly.

Several systematic reviews and meta-analyses were conducted on the diagnostic accuracy of IVC indexes to predict FR. In 2012, Mandeville and Colebourn³³ conducted a meta-analysis to evaluate the use of TTE in assessing dynamic markers of preload to predict FR in critically ill adult patients. They used broader selection criteria. Therefore, their study set was heterogeneous, including studies evaluating different measurement methods for FR, including IVC indexes. They concluded that meta-analysis cannot be performed because of the heterogeneity of the studies, as expected.³³ In 2014, Zhang et al.³⁴ conducted a meta-analysis focused on the diagnostic accuracy of the respiratory variation in IVC diameter as measured by bedside USG in predicting FR in critically ill patients. They included eight studies involving 235 patients, which comprised 6 of the 21 studies in our set. They concluded that the IVC diameter measured using USG is of great value in predicting FR, particularly in patients on controlled MV and in patients resuscitated with colloids. However, their study was limited by the small sample sizes of the included studies. In 2017, Long et al.³⁵ updated the systematic review to 17 studies involving 533 patients. They reported the pooled sensitivity and specificity of 0.63 (95% CI = 0.56–0.69) and 0.73 (95% CI = 0.67–0.78), respectively, with a pooled area under the receiver operating characteristic (AUROC) of 0.79. They conducted a subgroup analysis and found that respiratory variation in IVC diameter was a better predictor of FR in MV patients. In 2018, Si et al.³⁶ performed a similar meta-analysis by focusing on patients receiving MV. They included 12 studies involving 753 patients and performed subgroup analysis in the patient group ventilated with tidal volume (TV) ⩾8 mL/kg and PEEP ⩽5 cmH₂O. They reported the AUC of the SROC curve of the IVC diameter to predict FR in all patients on MV of 0.85 (95% CI = 0.81–0.86), sensitivity of 0.73 (95% CI = 0.60–0.84) and specificity of 0.82 (95% CI = 0.69–0.91). They concluded that the respiratory variability of the IVC diameter had limited ability for predicting FR in distinct ventilator settings, especially in patients with TV <8 mL/kg or PEEP >5 cmH₂O, and suggested that intensivists must be cautious when using respiratory variability of IVC diameter in those patients. In another meta-analysis published in 2018, Huang et al.³⁷ focused on patients with circulatory shock receiving MV for the first time. They included six studies involving 603 patients. They concluded that the respiratory variability of IVC diameter performed moderately well in predicting FR with an AUC of SROC, sensitivity and specificity of 0.82 (95% CI = 0.79–0.85), 0.69 (95% CI = 0.51–0.83) and 0.80 (95% CI = 0.66–0.89), respectively. Currently, Orso et al.³² evaluated the accuracy of the ‘caval index’ assessed using USG in predicting FR and included 20 studies involving 1709 cases. They reported that the pooled AUC, logarithmic diagnostic odds ratio, sensitivity and specificity were 0.71 (95% CI = 0.46–0.83), 2.02 (95% CI = 1.29–2.89), 0.71 (95% CI = 0.62–0.80) and 0.75 (95% CI = 0.64–0.85), respectively.³² The studies included in their meta-analysis were significantly different in their reference standard.

The pooled diagnostic utility estimates in this meta-analysis were quite similar to the previous studies. Our findings of the pooled positive and negative DLR of 3.86 and 0.34 suggest that respiratory variation in IVC diameter had moderate diagnostic utility for predicting the presence or absence of FR in adults.

We found that the diagnostic utility of the respiratory variability of IVC diameter for predicting FR was lower but not statistically significant in MV patients compared with those with spontaneous breathing. IVCc cannot be used for MV patients; therefore, studies on spontaneously breathing patients exclusively used the IVCc index. However, IVCd and IVCv were used for MV patients. Therefore, it was unclear if this difference was related to the index, methodological bias or real difference. This was similar to the findings of the latest review by Orso et al. in 2018.³² They proposed the lower extent of the changes in IVC diameter in MV patients as a reason for more approximation errors as an explanation.

We tried to decrease the heterogeneity of studies using explicit criteria for the selection of studies. We excluded studies on pregnant, paediatric or cancer patients, on dialysis patients or healthy volunteers, and in prehospital settings, which were proposed as possible covariates in previous studies. However, we still observed considerable heterogeneity among the reported diagnostic utility metrics of the included studies. The clinical variability of the critically ill patients is high; therefore, final considerations in this study may be unreliable, although we used robust statistical methods (particularly, a Bayesian method through a hierarchical approach) to overcome this limitation.

Conclusion

This meta-analysis showed that the respiratory variation of the IVC diameter had moderate diagnostic utility for predicting FR, regardless of the equation used to calculate the IVCv, with pooled positive and negative DLR of 3.86 and 0.34, respectively.

Footnotes

Acknowledgements

No assistance in the preparation of this article is to be declared.

Author contributions

E.U.A. contributed to the conceptualisation, methodology, investigation, data curation, writing – original draft, writing – review and editing software, supervision and project administration. H.A. contributed to the methodology, formal analysis, investigation, resources, data curation, writing – original draft, writing – review and editing software and visualisation.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship and/or publication of this article.

Availability of data and materials

The data sets generated and/or analysed during this study are available from the corresponding author.

Informed consent

Informed consent is not needed for systematic reviews.

Ethical approval

Ethical board approval is not needed for systematic reviews.

Human rights

This systematic review is performed under respectful conditions for human rights.

ORCID iDs

Ebru Unal Akoglu

Haldun Akoglu

References

Monnet

Teboul

. Prediction of fluid responsiveness in spontaneously breathing patients. Ann Transl Med 2020; 8: 790–790.

Bentzer

Griesdale

Boyd

, et al. Will this hemodynamically unstable patient respond to a bolus of intravenous fluids. JAMA 2016; 316: 1298–1309.

Salameh

Bossuyt

McGrath

, et al. Preferred reporting items for systematic review and meta-analysis of diagnostic test accuracy studies (PRISMA-DTA): explanation, elaboration, and checklist. BMJ 2020; 370: m2632.

Cooper

Hedges

Valentine

, et al. The handbook of research synthesis and meta-analysis. New York: Russell Sage Foundation, 2009.

Group CSDTM. Cochrane handbook for systematic reviews of diagnostic test accuracy. The Cochrane Collaboration, https://methods.cochrane.org/sdt/handbook-dta-reviews

Ouzzani

Hammady

Fedorowicz

, et al. Rayyan-a web and mobile app for systematic reviews. Syst Rev 2016; 5: 210–10.

Schueler

Schuetz

Dewey

. The revised QUADAS-2 tool. Ann Intern Med 2012; 156: 323–author reply 323–4.

Whiting

Rutjes

AWS

Westwood

, et al. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med 2011; 155: 529–536.

McGuinness

Higgins

JPT

. Risk-of-bias VISualization (robvis): an R package and Shiny web app for visualizing risk-of-bias assessments. Res Synth Methods 2021; 12: 55–61.

10.

Reitsma

Glas

Rutjes

AWS

, et al. Bivariate analysis of sensitivity and specificity produces informative summary measures in diagnostic reviews. J Clin Epidemiol 2005; 58: 982–990.

11.

Barbier

Loubières

Schmit

, et al. Respiratory changes in inferior vena cava diameter are helpful in predicting fluid responsiveness in ventilated septic patients. Intensive Care Med 2004; 30(9): 1740–1746.

12.

Charbonneau

Riu

Faron

, et al. Predicting preload responsiveness using simultaneous recordings of inferior and superior vena cavae diameters. Crit Care 2014; 18: 473–479.

13.

de Oliveira

Freitas

Ladeira

, et al. Comparison between respiratory changes in the inferior vena cava diameter and pulse pressure variation to predict fluid responsiveness in postoperative patients. J Crit Care 2016; 34: 46–49.

14.

Jiang

, et al. Exploring the best predictors of fluid responsiveness in patients with septic shock. Am J Emerg Med 2016; 35: 1258–1261.

15.

Machare-Delgado

Decaro

Marik

. Inferior vena cava variation compared to pulse contour analysis as predictors of fluid responsiveness: a prospective cohort study. J Intensive Care Med 2017; 26: 116–124.

16.

Moretti

Pizzi

. Inferior vena cava distensibility as a predictor of fluid responsiveness in patients with subarachnoid hemorrhage. Neurocrit Care 2011; 13: 3–9.

17.

Sobczyk

Nycz

Andruszkiewicz

, et al. Ultrasonographic caval indices do not significantly contribute to predicting fluid responsiveness immediately after coronary artery bypass grafting when compared to passive leg raising. Cardiovasc Ultrasound 2010; 14: 23–28.

18.

Theerawit

Morasert

Sutherasan

. Inferior vena cava diameter variation compared with pulse pressure variation as predictors of fluid responsiveness in patients with sepsis. J Crit Care 2016; 36: 246–251.

19.

Vignon

Repessé

Bégot

, et al. Comparison of echocardiographic indices used to predict fluid responsiveness in ventilated patients. Am J Respir Crit Care Med 2017; 195: 1022–1032.

20.

Yao

Liu

Sun

, et al. The value of the inferior vena cava area distensibility index and its diameter ratio for predicting fluid responsiveness in mechanically ventilated patients. Shock 2019; 52(1): 37–42.

21.

Zhang

Chen

, et al. Respiratory variations of inferior vena cava fail to predict fluid responsiveness in mechanically ventilated patients with isolated left ventricular dysfunction. Ann Intensive Care 2019; 9: 113–119.

22.

Airapetian

Maizel

Alyamani

, et al. Does inferior vena cava respiratory variability predict fluid responsiveness in spontaneously breathing patients? Crit Care 2015; 19: 400–408.

23.

Bortolotti

Colling

Colas

, et al. Respiratory changes of the inferior vena cava diameter predict fluid responsiveness in spontaneously breathing patients with cardiac arrhythmias. Ann Intensive Care 2018; 8: 79–12.

24.

Corl

George

Romanoff

, et al. Inferior vena cava collapsibility detects fluid responsiveness among spontaneously breathing critically-ill patients. J Crit Care 2017; 41: 130–137.

25.

Corl

Azab

Nayeemuddin

, et al. Performance of a 25% inferior vena cava collapsibility in detecting fluid responsiveness when assessed by novice versus expert physician sonologists. J Intensive Care Med 2020; 35: 1520–1528.

26.

Lanspa

Grissom

Hirshberg

, et al. Applying dynamic parameters to predict hemodynamic response to volume expansion in spontaneously breathing patients with septic shock. Shock 2013; 39: 155–160.

27.

McGregor

Sharma

Gupta

, et al. Emergency department non-invasive cardiac output study (EDNICO): an accuracy study. Scand J Trauma Resusc Emerg Med 2020; 28: 8–9.

28.

Muller

Bobbia

Toumi

, et al. Respiratory variations of inferior vena cava diameter to predict fluid responsiveness in spontaneously breathing patients with acute circulatory failure: need for a cautious use. Crit Care 2012; 16: R188.

29.

Preau

Bortolotti

Colling

, et al. Diagnostic accuracy of the inferior vena cava collapsibility to predict fluid responsiveness in spontaneously breathing patients with sepsis and acute circulatory failure. Crit Care Med 2017; 45(3): e290–e297.

30.

Feissel

Michard

Faller

, et al. The respiratory variation in inferior vena cava diameter as a guide to fluid therapy. Intensive Care Med 2004; 30(9): 1834–1837.

31.

Hao

Yang

, et al. Internal jugular vein variability predicts fluid responsiveness in cardiac surgical patients with mechanical ventilation. Ann Intensive Care 2018; 8: 6–9.

32.

Orso

Paoli

Piani

, et al. Accuracy of ultrasonographic measurements of inferior vena cava to determine fluid responsiveness: a systematic review and meta-analysis. J Intensive Care Med 2020; 35: 354–363.

33.

Mandeville

Colebourn

. Can transthoracic echocardiography be used to predict fluid responsiveness in the critically ill patient? A systematic review. Crit Care Res Pract 2012; 2012: 513480.

34.

Zhang

Luan

Zhu

, et al. Does ultrasonographic measurement of the inferior vena cava diameter correlate with central venous pressure in the assessment of intravascular volume in patients undergoing gastrointestinal surgery? J Surg Res 2014; 191: 339–343.

35.

Long

Oakley

Duke

, et al. Does respiratory variation in inferior vena cava diameter predict fluid responsiveness: a systematic review and meta-analysis. Shock 2017; 47: 550–559.

36.

Liu

, et al. Does respiratory variation in inferior vena cava diameter predict fluid responsiveness in mechanically ventilated patients? A systematic review and meta-analysis. Anesth Analg 2018; 127: 1157–1164.

37.

Huang

Shen

Liu

, et al. Value of variation index of inferior vena cava diameter in predicting fluid responsiveness in patients with circulatory shock receiving mechanical ventilation: a systematic review and meta-analysis. Crit Care 2018; 22: 204–207.

Does respiratory variation in inferior vena cava diameter predict fluid responsiveness in adult patients? A systematic review and meta-analysis of diagnostic accuracy studies

Abstract

Objectives:

Methods:

Results:

Conclusion:

Keywords

Introduction

Methods

Criteria for study selection

Types of studies

Participants

Index tests

Target condition and reference standards

Search methods for identification of studies

Electronic search

Searching other resources

Data collection and analysis

Study eligibility and quality assessment

Data extraction and management

Assessment of methodological quality

Statistical analysis and data synthesis

Investigations of heterogeneity

Sensitivity analysis

Assessment of reporting bias

Results

Search results

The methodological quality of the included studies

Findings

Summary of main results

Discussion

Conclusion

Footnotes

Acknowledgements

Author contributions

Declaration of conflicting interests

Funding

Availability of data and materials

Informed consent

Ethical approval

Human rights

ORCID iDs

References