The effects of static and dynamic measurements using transpulmonary thermodilution devices on fluid therapy in septic shock: A systematic review

Abstract

Transpulmonary thermodilution devices have been widely shown to be accurate in septic shock patients in assessing fluid responsiveness. We conducted a systematic review to assess the relationship between fluid therapy protocols guided by transpulmonary thermodilution devices on fluid balance and the amount of intravenous fluid used in septic shock. We searched MEDLINE, Embase and The Cochrane Library. Studies were eligible for inclusion if they were prospective, parallel trials that were conducted in an intensive care setting in patients with septic shock. The comparator group was either central venous pressure, early goal-directed therapy or pulmonary artery occlusion pressure. Studies assessing only the accuracy of fluid responsiveness prediction by transpulmonary thermodilution devices were excluded. Two reviewers independently performed the search, extracted data and assessed the bias of each study. In total 27 full-text articles were identified for eligibility; of these, nine studies were identified for inclusion in the systematic review. Three of these trials used dynamic parameters derived from transpulmonary thermodilution devices and six used primarily static parameters to guide fluid therapy. There was evidence for a significant reduction in positive fluid balance in four out of the nine studies. From the available studies, the results suggest the benefit of transpulmonary thermodilution monitoring in the septic shock population with regard to reducing positive fluid balance is seen when the devices are utilised for at least 72 hours. Both dynamic and static parameters derived from transpulmonary thermodilution devices appear to lead to a reduction in positive fluid balance in septic shock patients compared to measurements of central venous pressure and early goal-directed therapy.

Keywords

Septic shock haemodynamic monitoring fluid responsiveness cardiac output monitoring transpulmonary thermodilution

Introduction

Optimal fluid management in septic shock patients is unclear and consensus on the best approach has shifted greatly in the past 20 years.^1–3 Multiple trials have demonstrated that fluid therapy in sepsis is highly complex and needs to be tailored to the individual patient rather than relying on static pressure–based goal therapy.^4–6 To this end, advanced haemodynamic monitoring and dynamic assessment of how to best determine if a patient is fluid responsive has been well studied.^7–9 Common dynamic parameters are the change in cardiac output/index following a fluid challenge or passive leg raise (PLR) and stroke volume variation. Static parameters include pressure-based measurements including central venous pressure (CVP) and pulmonary capillary wedge pressure (PCWP), and volumetric measurements including extravascular lung water index (EVLWI), intrathoracic blood volume index (ITBVI) and global end-diastolic index (GEDI). The benefit of advanced haemodynamic monitoring over CVP or other pressure-based measurements has been well shown for reducing mortality, duration of length of stay (LOS) and duration of mechanical ventilation (MV) in mixed intensive care unit (ICU) populations.^10,11 The main problem with extrapolating these results to septic shock patients is these reviews have included non-calibrated devices and evidence suggests these devices are not able to accurately track the rapidly changing haemodynamics of septic patients when compared to either a pulmonary artery catheter (PAC) or transpulmonary thermodilution (TPTD) devices.^12,13 Additionally, most studies regarding advanced haemodynamic monitoring and septic shock are concerned with the accuracy of a device’s ability to judge if a patient is fluid responsive and do not address whether the ability to determine a patient’s fluid responsiveness leads to any meaningful change in decisions around volume of intravenous (IV) fluid used in the resuscitation and maintenance of the septic patient.

The underlying assumption of these studies comparing the accuracy of advanced haemodynamic to pressure-based measurements is that if dynamic monitoring is more accurate in judging fluid responsiveness, meaningful outcomes with regard to a reduction in positive fluid balance should be seen when these devices are used. To this end, this systematic review was conducted to compare IV volume therapy and clinical outcomes when dynamic and/or static TPTD parameters are used in comparison to clinical examination and pressure-based measurements.

Methods

Protocol and eligibility criteria

This systematic review was prospectively registered with PROSPERO (CRD42018106744). Studies eligible for inclusion were conducted in ICU patients with a diagnosis of septic shock, prospective in nature and included a study arm with allocation to fluid therapy guided by a TPTD device requiring intermittent calibration (either the Pulse index Contour Cardiac Output (PiCCO) device or EV1000 with VolumeView). The comparator group could be any other modality to assess fluid responsiveness including early goal-directed therapy (EGDT), CVP or other haemodynamic monitoring tool. Only studies that evaluated the effect of TPTD devices on total IV fluid use or clinical outcomes were included; studies only assessing the ability of the devices to predict fluid responsiveness were excluded. The primary outcome assessed could include IV fluid use or fluid balance, with no limitation on follow-up time. Secondary outcomes included duration of MV, LOS and mortality. Only peer-reviewed published studies were eligible for inclusion and there was no language restriction. Retrospective studies, reviews, abstracts and case reports were excluded.

Search and study selection

Two authors (TS and YH) searched MEDLINE, Embase and The Cochrane Library from the respective inception dates of the databases to May 2019. A selection of the keywords used in the search strategy were: ‘Fluid responsiveness’, ‘Fluid balance’ or ‘Fluid replacement’, ‘Septic shock’, ‘Sepsis’, ‘PiCCO’, ‘EV1000’, ‘Pulse pressure variation’, ‘Stroke volume’ and ‘Passive leg raise’. A full search strategy for each database can be found in Supplemental data 1. The bibliographies of included papers and relevant review articles were also scanned for additional articles. A two-step process for identifying studies was employed; initially TS and YH identified all titles/abstracts that potentially fulfilled the inclusion criteria. Full texts were then assessed for eligibility; if there was disagreement between the two assessors, a third assessor (SO) made the final decision.

Data collection and items

The following data were extracted by TS and YH to a pre-designed form: number of patients, country, inclusion/exclusion criteria, comparator protocol, and protocol for TPTD, age, gender and illness severity scores including Sequential Organ Failure Assessment and Acute Physiology and Chronic Health Evaluation scores. The primary outcome of fluid balance and IV fluid expansion volumes were recorded for each available timepoint reported by the individual study. Secondary outcomes were outcomes related to MV, LOS and mortality, either in the ICU or hospital.

Risk of bias in individual studies

The risk of bias for the included randomised controlled trials (RCTs) was evaluated by the Cochrane Risk of Bias tool and prospective cohort trials were assessed by ROBINS-1.

Data analysis

A meta-analysis was deemed inappropriate due to the different devices used in the comparator group, outcomes measured in each study and different endpoints with regard to time frame.

As TPTD devices can either rely on volumetric or dynamic fluid assessments, included studies were split into two groups for primary outcome analysis: those that primarily used dynamic methods and those that primarily used static TPTD parameters. Studies were defined as dynamic method studies if the fluid protocol focused on IV therapy guided by judging fluid responsiveness by dynamic assessment whereas static method studies were those that guided fluid therapy based on volumetric endpoints.

To reduce the heterogeneity of the fluid outcomes, where available 72-hour fluid balances have been reported. When 72-hour fluid balances were not available, the closest fluid balance in timing was reported. For individual study outcome data, including primary and secondary outcomes, either the mean (standard deviation) or median (interquartile range) has been used in this systematic review based on the original study’s reporting method.

Although a formal meta-analysis was not possible due to heterogeneity, simple pooled analysis of primary and secondary outcomes has been performed along with a narrative summary of the individual studies. The Grading of Recommendations Assessment, Development, and Evaluation (GRADE) approach was applied to the dynamic TPTD compared to static measurements and the static TPTD compared to static measurement subgroups to provide an overall assessment of the evidence relating to both the primary and secondary outcomes. The summary of findings table was generated using the GRADEpro software.¹⁴ Supplementary data file 2 contains justifications for assessment of the evidence ratings.

Results

Characteristics of included studies

As shown in Figure 1, from a total of 8162 citations retrieved, nine studies were selected for inclusion.^15–23 The nine studies were all prospective, with eight being RCTs and one being a prospective cohort. Six of the studies were published in English and three in Chinese. All trials had septic shock patients as their primary population of interest with three exceptions: Yu et al. (2017) limited their study population to chronic obstructive pulmonary disease (COPD) patients in septic shock, Trof et al. included septic shock and all other forms of shock and Zhang et al. included both acute respiratory distress syndrome (ARDS) and septic shock patients in their trial. There were three trials that used a dynamic TPTD protocol and six trials that used a protocol based on static TPTD measurements. Five of the comparator groups used CVP pressures, three groups used EGDT based on CVP measurements and one group used PCWP to guide fluid therapy.

Figure 1.

PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) flow chart of study inclusion.

A total of 813 septic shock patients were evaluated from the nine included studies. All nine studies used the PiCCO device as the TPTD device. The characteristics of the included studies are shown in Table 1.

Table 1.

Characteristics of included studies.

Study	Year/Country	Design/number of patients	Population	Time frame of eligibility	Age (years)	SOFA/APACHE II	Patients on mechanical ventilation at time of inclusion	Comparator	Outcomes assessed
	Studies which included TPTD dynamic parameters
Richard et al.¹⁵	2015/France	Single centre RCT/60	18-years-old or older patients in septic shock	Within 12 hours of onset of septic shock	64.5 (22)	10.5 (14) /N/A	77%	CVP	Intravascular volume expansion at 72 hours Duration of MVICU LOSMortality
Liu et al.¹⁶	2016/China	Single centre RCT/50	Septic shock patients who were mechanically ventilated	On admission to ICU	46.0 (8)	N/A / 18.5 (6)	100%	EGDT	Fluid balance at 48 hours
Trof et al. ^17,a	2012/Netherlands	Single centre RCT/120 (72 septic shock patients)	18- to 80-year-old shock patients on MV	At onset of shock but only if expected to stay longer than 48 hours in ICU	67 (17)	9.5 (6) / 26.5 (14)	100%	PAC	Fluid balance at 72 hoursFluid infusion 72 hours Duration of MVICU LOSMortality
	Trials that only used TPTD static parameters (GEDI/EVLWI/ITBVI)
Zhang et al.¹⁸	2015/China	Multicentre RCT/350	18-years-old or older patients in septic shock and/or who met ARDS	Within 24 hours of admission to ICU	63.5 (22)	9.5 (5) / 26.4 (14)	N/A	CVP	Fluid balance at 72 hoursICU LOSMortality
Xu, Y et al.¹⁹	2011/China	Single centre prospective cohort/33	All patients with septic shock	On admission to ICU	67.3 (15)	9.3 (4) / 24.0 (7)	N/A	CVP	Fluid input at 72 hoursFluid balance at 72 hours
Lu et al. ²⁰	2014/China	Single centre RCT/82	18-years-old or older patients with severe sepsis or septic shock	On admission to ICU	59.9 (18)	18.8 (7) / 28.3 (10)	82%	EGDT	Fluid input at 72 hoursDuration of MVICU LOSMortality
Yu et al. ²¹	2017/China	Single centre RCT/71	Septic shock patients with known COPD and on MV	On admission to ICU	70.5 (6)	N/A / 20.0 (5)	100%	CVP	24-hour fluid infusion totalsDuration of MVICU LOSMortality
Yao et al. ²²	2017/China	Single centre RCT/88	Septic shock patients	Not stated	63.0 (11)	N/A / N/A	N/A	CVP	Fluid balance (unclear time)Duration of MVICU LOS
Ma et al.²³	2017/China	Single centre RCT/40	Septic shock patients older than 65	On admission to ICU	77.3 (7)	8.6 (2) / 22.9 (4)	100%	EGDT	Fluid balance at 48 hoursDuration of MVICU LOSMortality

^aTrof separated all data analysis into septic shock and non-septic shock. All data in this systematic review are from the septic shock population and not the overall study population.

SOFA: sequential organ failure assessment; APACHE II: acute physiology and chronic health evaluation II; RCT: randomised controlled trial; MV: mechanical ventilation: COPD: chronic obstructive pulmonary disease; CVP: central venous pressure; EGDT: early goal-directed therapy; LOS: length of stay; TPTD: transpulmonary thermodilution; ICU: intensive care unit; GEDI: global end-diastolic index; EVLWI: extravascular lung water index; ITBVI: intrathoracic blood volume index; PAC: pulmonary artery catheter.

Data represented as mean (standard deviation).

With respect to bias, only one trial was assessed to have a low risk of bias, all other trials had a moderate, high/serious or unclear risk of bias with the most common risk of bias due to lack of blinding (Table 2).

Table 2.

Assessment of individual trial bias.

Papers (randomised controlled trials)	Random sequence generation	Allocation concealment	Blinding of participants and personnel	Blinding of outcome assessment	Incomplete outcome data	Selective reporting bias	Overall bias
Richard et al.¹⁵	Low risk	Low risk	High risk	Low risk	Low risk	Low risk	High risk
Liu et al. ¹⁶	Unclear risk	Unclear risk	High risk	High risk	Low risk	Unclear risk	High risk
Trof et al.¹⁷	Low risk	Low risk	High risk	High risk	Low risk	Unclear risk	High risk
Zhang et al.¹⁸	Low risk	Low risk	Low risk	Low risk	Low risk	Low risk	Low risk
Lu et al.²⁰	Low risk	Unclear risk	High risk	Unclear risk	Low risk	Unclear risk	High risk
Yu et al.²¹	Low risk	Unclear risk	High risk	High risk	Low risk	Unclear risk	High risk
Yao et al.²²	Unclear risk	Unclear risk	High risk	High risk	Low risk	Unclear risk	High risk
Ma et al.²³	Low risk	Unclear risk	High risk	Unclear risk	Low risk	Unclear risk	High risk
Papers (prospective cohorts)	Confounding	Selection of participants	Classification of interventions	Deviations from intended intervention	Incomplete data	Measurement of outcomes	Selection of reported result	Overall B bias
Xu, Y et al.¹⁹	Serious risk	Low risk	Low risk	Low risk	Low risk	Moderate risk	Moderate risk	Serious risk

For the nine randomised controlled trials (RCTs) included in this systematic review, the degree of bias was assessed by the Cochrane Risk of Bias tool and the one prospective cohort trial was assessed for bias by the ROBINS-1 tool.

Effect of TPTD on fluid therapy and clinical outcomes

The individual results of fluid-based outcomes for the three trials that used dynamic parameters measured by the TPTD are demonstrated in Table 3. The individual results of fluid-based outcomes for the six trials that used static parameters measured by the TPTD are demonstrated in Table 4. Overall, the effect of using TPTD monitoring in comparison to pressure-based measurements on fluid balance in septic shock is mixed and appears dependent on the intervention that TPTD is compared to and the time frame of the study more than the TPTD protocol used. Studies that reported at 6 or 24 hours either found no difference in fluid balance or infusion totals between TPTD and pressure measurements²³ or found an increase in fluid use in the TPTD group.^20,21 In contrast, studies that reported at 48 and 72 hours either found no significant difference^15,17,19 or a reduction in positive fluid balance for the TPTD group.^16,23 Finally, only one trial reported to seven days and found a less positive fluid balance on day seven for the TPTD group.¹⁸ In the seven trials comparing TPTD to CVP or EGDT based primarily on CVP, four demonstrated there was a significant reduction in the positive fluid balance when TPTD was used. In comparison, in the one trial that compared TPTD to PCWP, there was no difference in fluid balance.¹⁷ Assessing the effect of TPTD protocols based on dynamic compared to static measures, when dynamic TPTD studies were pooled for fluid balance, there was a moderate level of evidence to show a less positive fluid balance compared to pressure-based measurements. When static TPTD studies were pooled for fluid balance, there was a low level of evidence to show a reduction in positive fluid balance compared to pressure-based measurements.

Table 3.

Fluid-based outcomes in studies using dynamic transpulmonary thermodilution measures.

Trial	Transpulmonary thermodilution parameter	Early goal-directed therapy or central venous pressure parameters	Endpoints used	Fluid outcomes
Richard et al.¹⁵	Patient on MV with TV >7 ml/kg-> PPV ≥13% triggered a 500 ml bolus.ORIf patient was on MV with TV ≤ 7 ml/kg-> a PLR was performed and a change in stroke volume of 10% triggered a 500 ml bolus	CVP < 8 mmHg was used to guide further fluid resuscitation until CVP >8 mmHg	Fluid balance for ICU admission72-hour intravascular volume expansion	No significant difference in average fluid balance over an ICU admission (CVP: 1749 (1362) ml vs TPTD: 888 (2663) ml; P = 0.680).TPTD: 2850 (2600) ml vs CVP: 2900 (1625–5775) ml; P = 0.650
Liu et al.¹⁶	If GEDI <680 ml/m²/ITBVI <850 ml/m² and stroke volume variation ≥10% ->fluid resuscitation commenced until all parameters were within normal range	EGDT: If CVP was < 8 mmHg, then fluid resuscitation commenced until CVP 8–12 mmHg	48-hour fluid balance	48-hour fluid balance: TPTD: 7889 (1694) ml vs EGDT: 8427 (1328) ml; P < 0.05
Trof et al.¹⁷	EVLWI < 10 ml/kg -> fluid challenge and assess if fluid responsiveOREVLWI ≥10 ml/kg and GEDI > 850 ml/m², no further IV therapy	PCWP < 18 mmHg or < 20 mmHg if PEEP ≥10 cm -> fluid challenge. ORPCWP >18 mmHg or >20 mmHg if PEEP ≥10 cm, no further IV therapy	72-hour fluid balance72-hour fluid input	TPTD 1641 (2490) ml vs PAC 1979 (1435) ml; P = 0.490TPTD 3741 (1961) ml vs PAC 4130 (1804) ml; P = 0.390

TV: tidal volume; PPV: pulse pressure variation; PLR: passive leg raise; GEDI: global end-diastolic index; ITBVI: intrathoracic blood volume index; EVLWI: extravascular lung water index; PCWP: pulmonary capillary wedge pressure; PEEP: positive end-expiratory pressure; MV: mechanical ventilation; CVP: central venous pressure; TPTD: transpulmonary thermodilution; EGDT: early goal-directed therapy; PAC: pulmonary artery catheter. Data represented as mean (standard deviation) or median (interquartile range) where appropriate.

Table 4.

Fluid-based outcomes in studies utilising static transpulmonary thermodilution measures.

Trial	Transpulmonary thermodilution parameter	Early goal-directed therapy or central venous pressure parameters	Endpoints used	Fluid outcomes
Zhang et al.¹⁸	ITBVI < 850 ml/m² with an EVLWI <10 ml/kg -> fluid loading until ITBVI > 850 ml/m²	CVP < 8 mmHg-> 500 ml fluid bolus with aim of CVP 8–12 mmHg	Day 3 fluid balance	TPTD: 300 (1869) ml vs CVP 250 (1750) ml; P > 0.05
Xu, Y et al.¹⁹	Fluid resuscitation given to aim for an ITBVI 850–1000 ml/m²	CVP < 8 mmHg -> fluid resuscitation with aim of CVP 8–12 mmHg	72- hour fluid balance72-hour fluid input	TPTD: 3882 (889) ml vs CVP: 4703 (2813) ml; P>0.05TPTD: 9918 (137) ml vs CVP: 10529 (331) ml; P>0.05
Lu et al.²⁰	Fluid resuscitation given to aim for an ITBVI of 850–1000 ml/m²	EGDT with CVP <8 mmHg triggering fluid resuscitation	72-hour fluid balance	TPTD: 918 (315) ml vs EGDT: 1805 (420) ml; P=0.037
Yu et al.²¹	Fluid resuscitation to target a GEDI of 800 ml/m²	Fluid resuscitation to target a CVP of 12 mmHg	24-hour fluid infusion	TPTD: 6327.3 (266) ml vs CVP: 4761.4 (239) ml; P<0.05
Yao et al.²²	ITBVI < 850 ml/m² with an EVLWI < 7ml/kg to trigger fluid resuscitation to aim ITBVI> 1000 ml/m² and EVLWI 7–10 ml/kg	CVP <8 mmHg -> fluid resuscitation with aim of CVP 8–12 mmHg	Fluid balance (unclear duration)	TPTD: 1733 (253) ml vs CVP: 2135.4 (205) ml; P<0.05
Ma et al.²³	Fluid resuscitation given to aim for a GEDI of 650–800 ml/m²	EGDT with CVP < 8 mmHg triggering fluid resuscitation	48-hour fluid balance	TPTD: 3141.6 (1246) ml vs EGDT:4533.9 (1417) ml; P<0.05

GEDI: global end-diastolic index; ITBVI: intrathoracic blood volume index; EVLWI: extravascular lung water index; MV: mechanical ventilation; CVP: central venous pressure; TPTD: transpulmonary thermodilution; EGDT: early goal-directed therapy; PAC: pulmonary artery catheter.

Data represented as mean (standard deviation).

Studies that assessed the fluid infusion totals demonstrated that total IV fluid use was either the same^15,17,19 or increased²¹ in the TPTD group compared to the pressure-based comparator groups. When dynamic TPTD studies were pooled for fluid infusion totals, there was a low level of evidence to show TPTD had no effect on infusion totals. When static TPTD studies were pooled for fluid infusion totals, there was a low level of evidence to show that static TPTD monitoring increased the total volume of IV fluid infusion use compared to pressure-based measurements.

Finally, the individual outcomes of each trial on effect of LOS and duration of MV (seven trials), as well as the mortality outcomes (six trials) are shown in Table 5. There was moderate evidence that TPTD monitoring lowered the duration of MV compared to pressure-based measurements in septic shock. The GRADE system has been used to generate a summary of evidence table (Table 6), presenting the quality of evidence for all the above outcomes.

Table 5.

Clinical outcomes for transpulmonary thermodilution in septic shock.

Trial	Duration of mechanical ventilation (days)	Length of stay (days)	Mortality
Richard et al.¹⁵	TPTD: 14 (24) vs CVP: 8 (21); P= 0.35	TPTD: 14 (22) vs CVP: 10 (13); P = 0.55	28-day mortality: TPTD 23% vs CVP 47%; P = 0.10
Liu et al.¹⁶	Did not report	Did not report	Did not report
Trof et al.¹⁷	TPTD: 10 (5) vs PAC: 13 (14); P = 0.39	TPTD: 11 (17) vs PAC 15 (15); P = 0.30	28-day mortality: TPTD: 42% vs PAC 38%; P = 0.83
Zhang et al.¹⁸	TPTD: 6 (9) vs CVP 5.5 (9); P = 0.897	TPTD: 9 (8) vs CVP: 7.5 (11), P = 0.598	28-day mortality: TPTD 23% vs CVP 47%; P = 0.10
Xu et al.¹⁹	Did not report	Did not report	Did not report
Lu et al.²⁰	TPTD: 4.1 (1) vs EGDT: 6.0 (1); P = 0.001	TPTD: 7.1 (3) vs EGDT: 9.5 (3); P = 0.004	In-hospital mortality: TPTD 16.7％ vs EGDT 17.5%; P = 0.921
Yu et al.²¹	TPTD: 3.1 (1) vs CVP: 4.2 (2), P = 0.032	TPTD: 7.4 (2) vs CVP: 11.3 (2); P = 0.003	In ICU mortality: TPTD 20.6 % vs CVP: 27%; P = 0.862
Yao et al.²²	TPTD: 16.3 (5) vs CVP: 19.6 (10); P < 0.05	TPTD: 6.3 (2) vs CVP:10.4 (3); P < 0.05	Not reported
Ma et al.²³	TPTD: 3.7 (1) vs EGDT: 4.5 (1); P < 0.05	TPTD: 5.8 (2) vs EGDT: 7.2 (2); P < 0.05	ICU mortality: TPTD 25.0% vs EGDT 40.0%; P = 0.311

CVP: central venous pressure; TPTD: transpulmonary thermodilution; EGDT: early goal-directed therapy; PAC: pulmonary artery catheter; ICU: intensive care unit.

Data represented as mean (standard deviation).

Table 6.

Grading of Recommendations Assessment, Development, and Evaluation summary of findings for the use of transpulmonary thermodilution in septic shock.

	Certainty assessment							No. of patients
	No. of studies	Study design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Transpulmonary thermodilution measures	Control	Effect	Certainty
Intravenous fluid balance with dynamic transpulmonary thermodilution measures ^15–17	3	Randomised trials	Serious	Not serious	Not serious	Serious	Strong association	93	89	One of two studies comparing dynamic TPTD to CVP showed significant reduction in positive fluid balance.The one study comparing TPTD to PCWP showed no effect on fluid balance.Simple pooled analysis of the trials showed a mean 603.3 (978) ml lower in the TPTD group.	⨁⨁⨁◯MODERATE
Intravenous fluid infusion totals with dynamic transpulmonary thermodilution measures ^15,17	2	Randomised trials	Serious	Not serious	Not serious	Serious	None	84	82	The two trials which compared dynamic TPTD to CVP and PCWP showed no significant effect on fluid infusion totals.Simple pooled analysis of the trials showed a mean 312.8 (1370) ml lower in the TPTD group.	⨁⨁◯◯LOW
Intravenous fluid balance with static transpulmonary thermodilution measures ^{18–20,22,23}	5	Randomised trials	Serious	Serious	Serious	Not serious	Strong association	252	265	Three of the five studies comparing static TPTD to CVP showed a significant reduction in positive fluid balance. Simple pooled analysis of the trials showed a mean 502.2 (158) ml lower in the TPTD group.	⨁⨁◯◯LOW
Intravenous fluid infusion totals with static transpulmonary thermodilution measures ^19,21	2	1 randomised trials + 1 prospective cohort	Very serious	Serious	Not serious	Serious	Strong association	94	92	One of the two trials comparing static TPTD to CVP showed a significant increase in IV fluid use.Simple pooled analysis of the trials showed a mean 1497.8 (232) ml higher in the TPTD group.	⨁⨁◯◯LOW
Duration of mechanical ventilation with transpulmonary thermodilution ^{15,17,18,20–23}	7	Randomised trials + 1 prospective cohort	Serious	Serious	Serious	Not serious	None	376	387	Three out of the seven trials that assessed the effect of TPTD monitoring compared to either CVP or PCWP showed a lower duration of MV.Simple pooled analysis of the trials showed the duration of mechanical ventilation was 0.55 days (0.7) less in the TPTD group.	⨁◯◯◯VERY LOW
Length of stay with transpulmonary thermodilution ^{15,17,18,20–23}	7	Randomised trials + 1 prospective cohort	Not serious	Not serious	Serious	Serious	Strong association	376	387	Three out of the seven trials that assessed the effect of TPTD monitoring compared to either CVP or PCWP showed a lower duration of LOS.Simple pooled analysis of the trials showed a length of stay of 2.51 (1) days less in the TPTD group.	⨁⨁⨁◯MODERATE
Mortality with transpulmonary thermodilution ^{15,17,18,20,21,23}	6	Randomised trials + 1 prospective cohort	Serious	Serious	Serious	Serious	None	125/332 (37.7%)	142/343 (41.4%)	None of the included trials demonstrated a significant effect on mortality.Simple pooled analysis of the trials showed an OR 0.86 (0.6) for mortality in the TPTD group or 36 fewer deaths per 1,000 (147).	⨁◯◯◯VERY LOW

CVP: central venous pressure; TPTD: transpulmonary thermodilution; EGDT: early goal-directed therapy; PAC: pulmonary artery catheter; OR: odds ratio; IV: intravenous; MV: mechanical ventilation; PCWP: pulmonary capillary wedge pressure; LOS: length of stay.

Data represented as mean (standard deviation).

Discussion

In our systematic review, we found the use of TPTD devices led to an overall reduction in positive fluid balance in septic shock compared to CVP and EGDT. All trials in this systematic review utilising EGDT as the comparator group were reliant on static measurements, most commonly CVP targets. Pressure-based targets in fluid management in septic shock patient have been shown to be detrimental to fluid balances and patient outcomes, likely because of the inherent pitfalls in interpreting volume status from pressure measurements.^3,6,24

Interestingly, although both dynamic and static measures of fluid assessment with TPTD devices led to significantly less positive fluid balances, dynamic fluid assessments had no effect on the total IV fluid used and static measures resulted in an increase in total IV fluid administration. It is unclear why there is a trend towards a less positive fluid balance with these devices if they result in an increase in IV fluid therapy. This has been observed in other trials using non-calibrated devices in septic patients.²⁵ One possible reason is that the presence of advanced haemodynamic devices encourages the clinician to give fluid boluses to measure fluid responsiveness.²⁶ This may result in higher IV fluid infusions but changes the clinician’s decision-making in terms of strategies to limit the overall fluid balance.

The major limitation of this systematic review is the heterogeneity in the time frame of reporting on fluid outcomes. A prominent feature of trials linking fluid balance to mortality is that a prolonged large positive fluid balance at the end of day 2 to 4 following the onset of critical illness leads to the increase in mortality, not necessarily the volume given in the initial resuscitation period.^27,28 Few sepsis trials appreciate the significance of the phases of sepsis and will usually recruit all septic patients, irrespective of duration of sepsis, or use a surrogate time for the onset of sepsis (i.e. admission to the ICU). Based on the results of this systematic review, it appears the time frame of fluid endpoints has a large influence on the outcome and makes interpretation of pooled results difficult due to high levels of heterogeneity. The impact of this potential biphasic effect of fluids in septic shock is that it makes trials with different time frames extremely difficult to interpret in a systematic review. What is likely underestimated is the dynamic nature of sepsis and how quickly the patient may shift from one phase to the next. It has been demonstrated that the later fluids are initiated, the less positive effect they seem to have.²⁹ As illustrated by Zhong and Xue (2015), 85.7% of patients in septic shock were judged fluid responsive at the time of enrolment to the study but at 72 hours, only 28.6% patients were fluid responsive. This effect of changing fluid responsiveness may explain the time-related effect on fluid outcomes within each trial. Ma found no difference at 6 or 24 hours in fluid balance between GEDI/EVLWI protocol and EGDT but there was a significantly lower positive fluid balance at 48 hours for the GEDI/EVLWI group.²³ Zhang found no difference in fluid balance at 72 hours but demonstrated a significant reduction in positive fluid balance on day 7 for the TPTD group.¹⁸ Richard demonstrated no difference at 72 hours for fluid infusion between the two groups but when considering total fluids over the duration of admission with average LOS of 12 days, there was significantly less IV fluid use on a daily basis in the dynamic TPTD group compared to CVP.¹⁵ Trof showed higher fluid use in the first 24 hours for the TPTD group but less at 72 hours.¹⁷ The one trial, by Yu et al. (2017), which demonstrated increased fluid infusion totals using a TPTD device reported only the first 24 hours.²¹

It should be noted that not all TPTD directed therapy is equal in its accuracy to guide fluid therapy. Previous trials have demonstrated that dynamic parameters are far more accurate than static parameters in judging fluid responsiveness and therefore, accurate fluid administration.^30–32 Static volume-based measurements (GEDI/EVLWI) have numerous pitfalls in the septic shock population. In studies directly assessing the predictive ability of GEDI/ITBVI for fluid responsiveness, it has been shown to be no better than CVP.^33–35 There are other changes in the septic state that confound many of the static parameters. Septic cardiomyopathy changes, including global cardiac dilatation and left ventricular systolic impairment, likely impair the GEDI ability to guide fluids in the septic population and a standard GEDI of 680–800 ml/m² is almost certainly not appropriate for all septic patients.^35,36 Easley et al. (2009) showed the impairment of hypoxic pulmonary vasoconstriction, seen in some forms of sepsis, can cause an increase in the EVLWI without an actual increase in the amount of pulmonary oedema.³⁷ Although these limitations should indicate that dynamic assessments have a greater impact on fluid therapy, the evidence from this review suggests static and dynamic measures have relatively similar outcomes on fluid therapy.

Two additional major limitations that should be addressed in future trials are the lack of blinded randomised controlled trials assessing effect of fluid responsiveness on fluid balances and the complexity of dynamic measures protocols leading to lack of compliance.

There are few studies available to assess a TPTD’s effect on fluid therapy in septic shock; only nine studies were available for inclusion that focused on the overall fluid therapy and not the ability of the device to predict fluid responsiveness. Additionally, only one of the nine trials was blinded. Non-blinded study designs certainly lend themselves to a high degree of observer bias and Hawthorne effect, introducing a great degree of bias that weakens any result of these trials.

A common issue with dynamic fluid responsiveness guided fluid management is that despite their superior evidence for accuracy, they are driven by complex protocols. These complex processes lead to a high rate of non-compliance/adherence to the dynamic fluid protocols, likely leading to erroneous outcomes. Richard et al. (2015) reported 43% of patients in the dynamic group had at least one protocol violation.¹⁵ A possible reason for this non-adherence may be due to many clinicians being unfamiliar with dynamic fluid assessment techniques. As illustrated by Fischer et al. (2016), only 60% of senior ICU physicians could correctly apply pulse pressure variation (PPV); this fell to 17% when considering physicians who do not commonly use PPV.³⁸ This issue of non-compliance can be addressed and when education is provided, more profound effects of dynamic TPTD measurements are uncovered. In a trial conducted by Rameau et al. involving PLR and PiCCO monitoring in septic shock patients, once non-compliance was addressed, there was a significant fall in the average 48-hour fluid balance from 4.3 l to 3.1 l.³⁹

Based on the individual studies and GRADE assessment, there is a less clear picture for the effect of either dynamic or static TPTD protocols on LOS, duration of MV and mortality. Although the study protocols that relied on only volume-based measurements (GEDI and EVLWI) had weak evidence for improving the fluid balance compared to EGDT or CVP, they had similar effects on LOS and duration of MV as protocols that incorporated both dynamic and static parameters. This may indicate that although static parameters derived from TPTD should not be used to judge if a patient is fluid responsive, they are still highly useful in the management of the septic shock patient.^40–45

A strength of this systematic review is the focus of analysis on septic shock patients, as a frequent criticism of many ICU trials is that the largely heterogeneous group of patients in any given ICU may weaken results.^46,47 Sepsis is a diverse population; including other diseases to this systematic review would likely confound any meaningful conclusions to be drawn.

There are multiple limitations of this systematic review, which weaken any conclusions that can be derived about the effects of TPTD monitoring on septic shock patients. An inherent limitation of our paper due to its narrow focus is the limited number of trials available for inclusion in this systematic review and the overall small number of patients. There are multiple studies that demonstrate the effectiveness of advanced haemodynamic monitoring for improving fluid therapy in the ICU. However, these trials had to be excluded from this review, either because they focused on the non-septic population,^48,49 were in the postoperative population or utilised non-calibrated haemodynamic monitoring.⁵⁰ Another potential weakness in this systematic review is the failure to include other novel techniques, including peripheral arterial velocities and tidal volume challenges.^51,52 At the time of writing, these novel techniques do not yet have the level of evidence to determine whether they can be accurately applied in the septic shock population. However, these techniques are highly promising as they are less invasive than TPTD or PAC and easily applied, which therefore may reduce non-adherence to dynamic fluid protocols.

A final limitation of this review is not including PAC trials in this review article. Although both PACs and TPTDs have strong evidence supporting their accuracy in predicting fluid responsiveness in septic shock, TPTDs are preferred to PACs in many ICUs due to a perceived increased risk of PACs and assumed no added benefit over TPTDs.^53–56 The decision to exclude PACs from this review was made based on the evidence of decreasing PAC use in ICUs.⁵⁶ However, there is an increase in the use of PACs with growing recognition of their importance in septic shock patients with cardiac dysfunction, particularly involving the right ventricle. Although a TPTD is as good as a PAC device for tracking real-time haemodynamic changes in hyperdynamic states such as sepsis, it has limitations when right ventricular dysfunction or severe pulmonary hypertension are present and may be worse than a PAC for management of ARDS.^33,56,57 In the one trial that compared TPTD to PAC in this systematic review, there was no significant difference between fluid balance between the two groups in the septic shock population.¹⁷ However, the most effective and accurate method will likely be when TPTD or PAC is used in conjunction with other assessment methods to deliver a more comprehensive answer to the patient’s fluid status. A study by Vignon et al. (2018) assessing agreement of TPTD and transthoracic echocardiography found that echocardiography was able to identify situations where TPTD measurements were likely to be inaccurate.⁵⁸ This underlines that the best form of haemodynamic monitoring for fluid therapy in complex patients will likely be a combination of modalities and not one single device or measurement.

Conclusion

The evidence for the effect of TPTD devices on septic shock fluid therapy suggests that both dynamic and static measurements result in a significant reduction in positive fluid balance in the septic shock patient compared to pressure-based fluid protocols. However, the effect is likely underestimated for multiple reasons including non-adherence to protocols and inadequate study designs. Higher-quality trials are required in this field to properly assess the effect of haemodynamic monitoring on septic shock.

Supplemental Material

AIC893703 Supplemental material - Supplemental material for The effects of static and dynamic measurements using transpulmonary thermodilution devices on fluid therapy in septic shock: A systematic review

Supplemental material, AIC893703 Supplemental material for The effects of static and dynamic measurements using transpulmonary thermodilution devices on fluid therapy in septic shock: A systematic review by Timothy G Scully, Yifan Huang, Stephen Huang, Anthony S McLean and Sam R Orde in Anaesthesia and Intensive Care

Footnotes

Availability of data and materials

Complete search data and search histories are available from the corresponding author on request.

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.

ORCID iD

Timothy G Scully

Notes

Systematic review registration number (PROSPERO): CRD42018106744 ()

Supplemental material

Supplemental material for this article is available online.

References

Rivers

Nguyen

Havstad

, et al. Early goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med 2001; 345: 1368–1377.

Pittard

Huang

McLean

, et al. Association of positive fluid balance and mortality in sepsis and septic shock in an Australian cohort. Anaesth Intensive Care 2017; 45: 737–743

Marik

PE.

The demise of early goal-directed therapy for severe sepsis and septic shock.

Acta Anaesthesiol Scand 2015; 59: 561–567

Chen

Kollef

MH.

Targeted fluid minimization following initial resuscitation in septic shock: A pilot study.

Chest 2015; 148: 1462–1469

Alsous

Khamiees

DeGirolamo

, et al. Negative fluid balance predicts survival in patients with septic shock: A retrospective pilot study. Chest 2000; 117: 1749–1754

Kalil

Johnson

Lisco

, et al. Early goal-directed therapy for sepsis: A novel solution for discordant survival outcomes in clinical trials. Crit Care Med 2017; 45: 607–614.

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

Elwan

Roshdy

Elsharkawy

, et al. The haemodynamic dilemma in emergency care: Is fluid responsiveness the answer? A systematic review. Scand J Trauma Resusc Emerg Med 2017; 25: 25

Marik

Lemson

Fluid responsiveness: An evolution of our understanding.

Br J Anaesth 2014; 112: 617–620

10.

Ansari

Zochios

Falter

, et al. Physiological controversies and methods used to determine fluid responsiveness: A qualitative systematic review. Anaesthesia 2016; 71: 94–105

11.

Bednarczyk

Fridfinnson

Kumar

, et al. Incorporating dynamic assessment of fluid responsiveness into goal-directed therapy: A systematic review and meta-analysis. Crit Care Med 2017; 45: 1538–1545

12.

Sakka

Kozieras

Thuemer

, et al. Measurement of cardiac output: A comparison between transpulmonary thermodilution and uncalibrated pulse contour analysis. Br J Anaesth 2007; 99: 337–342

13.

Grensemann

Defosse

Wieland

, et al. Comparison of PulsioFlex(R) uncalibrated pulse contour method and a modified Fick principle with transpulmonary thermodilution measurements in critically ill patients. Anaesth Intensive Care 2016; 44: 484–490

14.

GRADEpro GDT: GRADEpro Guideline Development Tool [Software]. McMaster University dbEP, Inc.

15.

Richard JC, Bayle F, Bourdin G, et al. Preload dependence indices to titrate volume expansion during septic shock: A randomized controlled trial. Crit Care 2015; 19: 5.

16.

Liu

Xie

, et al. Passive leg raising predicts volume responsiveness in patients with septic shock. Chin J Surg 2011; 49: 44–48.

17.

Trof

Beishuizen

Cornet

, et al. Volume-limited versus pressure-limited hemodynamic management in septic and nonseptic shock. Crit Care Med 2012; 40: 1177–1185.

18.

Zhang

Qian

Effectiveness of treatment based on PiCCO parameters in critically ill patients with septic shock and/or acute respiratory distress syndrome: A randomized controlled trial.

Intensive Care Med 2015; 41: 444–451.

19.

Liu

, et al. Intrathoracic blood volume index as an indicator of fluid management in septic shock. [Chinese]. Chinese Crit Care Med 2011; 23: 462–426.

20.

Zheng

Lin

, et al. Clinical studies of surviving sepsis bundles according to PiCCO on septic shock patients. [Chinese]. Chinese Crit Care Med 2014; 26: 23–27.

21.

Zheng

Lin

, et al. Global end-diastolic volume index vs CVP goal-directed fluid resuscitation for COPD patients with septic shock: a randomized controlled trial. Am J Emerg Med 2017; 35: 101–105.

22.

Yao

Guan

Zhuang

, et al. Effects of fluid resuscitation under the guidance of PICCO on the immune function and inflammatory mediator in patients with septic shock. Biomed Res (Aligarh) 2017; 28: 9722–9725.

23.

Zhang

Wang

, et al. Effect of global end diastolic volume index guidance fluid resuscitation in elderly patients with septic shock. Chinese Crit Care Med 2017; 29: 486–490.

24.

Saugel

Ringmaier

Holzapfel

, et al. Physical examination, central venous pressure, and chest radiography for the prediction of transpulmonary thermodilution-derived hemodynamic parameters in critically ill patients: A prospective trial. J Crit Care 2011; 26: 402–410.

25.

Hou

Filbin

Napoli

, et al. Cardiac Output Monitoring Managing Intravenous Therapy (COMMIT) to treat emergency department patients with sepsis. Shock 2016; 46: 132–138.

26.

Scully

Grealy

McLean

, et al. Calibrated cardiac output monitoring versus standard care for fluid management in the shocked ICU patient: A pilot randomised controlled trial. J Intensive Care 2019; 7: 1.

27.

de Oliveira

Freitas

Ferreira

, et al. Positive fluid balance as a prognostic factor for mortality and acute kidney injury in severe sepsis and septic shock. J Crit Care 2015; 30: 97–101.

28.

Boyd

Forbes

Nakada

, et al. Fluid resuscitation in septic shock: A positive fluid balance and elevated central venous pressure are associated with increased mortality. Crit Care Med 2011; 39: 259–265.

29.

Kern

Shoemaker

WC.

Meta-analysis of hemodynamic optimization in high-risk patients.

Crit Care Med 2002; 30: 1686–1692.

30.

Bentzer

Griesdale

Boyd

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

31.

Kanova

Sevcik

Static and dynamic tests for management of volumotherapy. [Czech].

Anesteziologie a Intenzivni Medicina 2014; 25: 107–116.

32.

Cheng

Liu

, et al. Predicting stroke volume and arterial pressure fluid responsiveness in liver cirrhosis patients using dynamic preload variables. Eur J Anaesthesiol 2016; 33: 645–652.

33.

Monnet

Teboul

J-L.

Transpulmonary thermodilution: Advantages and limits.

Crit Care 2017; 21: 147.

34.

Muller

Louart

Bengler

, et al. The intrathoracic blood volume index as an indicator of fluid responsiveness in critically ill patients with acute circulatory failure: A comparison with central venous pressure. Anesth Analg 2008; 107: 607–613.

35.

Endo

Kushimoto

Yamanouchi

, et al. Limitations of global end-diastolic volume index as a parameter of cardiac preload in the early phase of severe sepsis: A subgroup analysis of a multicenter, prospective observational study. J Intensive Care 2013; 1: 11.

36.

Malbrain

De Potter

, et al. Global and right ventricular end-diastolic volumes correlate better with preload after correction for ejection fraction. Acta Anaesthesiol Scand 2010; 54: 622–631.

37.

Easley

Mulreany

Lancaster

, et al. Redistribution of pulmonary blood flow impacts thermodilution-based extravascular lung water measurements in a model of acute lung injury. Anesthesiol 2009; 111: 1065–1074.

38.

Fischer

Mahjoub

Boisselier

, et al. Arterial pulse pressure variation suitability in critical care: A French national survey. Anaesthesia Crit Care Pain Med 2015; 34: 23–28.

39.

Rameau

de With

Boerma

EC.

Passive leg raise testing effectively reduces fluid administration in septic shock after correction of non-compliance to test results.

Ann Intensive Care 2017; 7: 2.

40.

Caltabeloti

Monsel

Arbelot

, et al. Early fluid loading in acute respiratory distress syndrome with septic shock deteriorates lung aeration without impairing arterial oxygenation: A lung ultrasound observational study. Crit Care (London, England) 2014; 18: R91.

41.

Chew

Ihrman

During

, et al. Extravascular lung water index improves the diagnostic accuracy of lung injury in patients with shock. Crit Care 2012; 16: R1.

42.

Chung

F-T

Lin

H-C

Kuo

C-H

, et al. Extravascular lung water correlates multiorgan dysfunction syndrome and mortality in sepsis. PloS one 2010; 5: e15265.

43.

Diaz-Rubia

Ramos-Saez

Vazquez-Guillamet

, et al. Efficacy of an extravascular lung water-driven negative fluid balance protocol. Med Intensiva 2015; 39: 345–351.

44.

Jozwiak

Teboul

J-L

Monnet

Extravascular lung water in critical care: Recent advances and clinical applications.

Ann Intensive Care 2015; 5: 38.

45.

Saugel

Wildgruber

Staudt

, et al. Quantitative computed tomography in comparison with transpulmonary thermodilution for the estimation of pulmonary fluid status: A clinical study in critically ill patients. J Clin Monit Comput 2018; 33: 5–12.

46.

Harhay

Wagner

Ratcliffe

, et al. Outcomes and statistical power in adult critical care randomized trials. Am J Respir Crit Care Med 2014; 189: 1469–1478.

47.

Vincent

J-L

Marini

Pesenti

Do trials that report a neutral or negative treatment effect improve the care of critically ill patients? No.

Intensive Care Med 2018; 44: 1989–1991.

48.

Sanchez

Garcia-de-Lorenzo

Herrero

, et al. A protocol for resuscitation of severe burn patients guided by transpulmonary thermodilution and lactate levels: A 3-year prospective cohort study. Crit Care 2013; 17: R176.

49.

Sun

Zhang

, et al. The effects of fluid resuscitation according to PiCCO on the early stage of severe acute pancreatitis. Pancreatol Off J Int Assoc Pancreatol 2015; 15: 497–502.

50.

Latham

Bengtson

Satterwhite

, et al. Stroke volume guided resuscitation in severe sepsis and septic shock improves outcomes. J Crit Care 2017; 42: 42–46.

51.

Myatra

Prabu

Divatia

, et al. The changes in pulse pressure variation or stroke volume variation after a ‘tidal volume challenge’ reliably predict fluid responsiveness during low tidal volume ventilation. Crit Care Med 2017; 45: 415–421.

52.

Tapanwong N, Chuntupama P and Wacharasint P. Dynamic changes in esophageal doppler derived parameters by elevation of peep, predict fluid responsiveness in patients with septic shock. Respirol 2017; 22 (Supplement 3): 122.

53.

Harvey

Harrison

Singer

, et al. Assessment of the clinical effectiveness of pulmonary artery catheters in management of patients in intensive care (PAC-Man): A randomised controlled trial. Lancet 2005; 366: 472–477.

54.

Harvey

Stevens

Harrison

, et al. An evaluation of the clinical and cost-effectiveness of pulmonary artery catheters in patient management in intensive care: A systematic review and a randomised controlled trial. Health Technol Assess 2006; 10: iii–iv, ix–xi, 1–133

55.

Ikuta

Wang

Robinson

, et al. National trends in use and outcomes of pulmonary artery catheters among medicare beneficiaries, 1999–2013. JAMA Cardiol 2017; 2: 908–913.

56.

De Backer

Vincent

JL.

The pulmonary artery catheter: Is it still alive?

Curr Opin Crit Care 2018; 24: 204–208.

57.

Saugel

Vincent

JL.

Cardiac output monitoring: How to choose the optimal method for the individual patient.

Curr Opin Crit Care 2018; 24: 165–172.

58.

Vignon

Begot

Mari

, et al. Hemodynamic assessment of patients with septic shock using transpulmonary thermodilution and critical care echocardiography: A comparative study. Chest 2018; 153: 55–64.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.06 MB