Pulmonary vascular permeability index and global end-diastolic volume: are the data consistent in patients with femoral venous access for transpulmonary thermodilution: a prospective observational study

Appropriate fluid supply is a cornerstone of critical care. In addition to cardiac output CO, transpulmonary thermodilution (TPTD) provides extravascular lung water (EVLW) reflecting pulmonary edema and the preload parameter global end-diastolic volume (GEDV) [1‐6]. Increased EVLW is associated to mortality [3, 7‐11]. GEDV and its changes are associated to CO [6] and algorithms based on GEDV have been shown to improve outcome [12]. Moreover, the relation of EVLW and pulmonary blood volume (PBV) might be useful to differentiate the etiology of pulmonary edema [13, 14]. PBV can be determined using double-indicator TPTD technique [13]. Single indicator TPTD is much more common in clinical routine and estimates PBV as about 25% of GEDV [14]. EVLW/PBV-ratio has been termed pulmonary vascular permeability index “PVPI”. Several studies demonstrated that higher values of PVPI are associated to pulmonary edema with increased permeability of the alveolo-capillary barrier [14]. Increased PVPI has been found in patients with ARDS resulting from pulmonary (e.g. pneumonia) or secondary (e.g. sepsis) pulmonary impairment. By contrast hydrostatic pulmonary edema due to congestive heart failure usually does not result in marked increases of PVPI since EVLW and PBV are increased to a similar extent [13, 14].

Although TPTD has become part of clinical routine, several pitfalls have to be kept in mind. Two recent studies demonstrated marked overestimation of GEDV in case of performing TPTD indicator injection using a femoral venous access due to the additional volume of vena cava inferior (VCI) participating in TPTD [15, 16]; Figure 1).

The authors suggested a correction formula for GEDV indexed to predicted body surface area (GEDVI) based on data derived from femoral TPTD and biometric information. As a consequence of these data one the manufacturer of the PiCCO device (Pulsion Medical Systems, Feldkirchen, Germany) implemented a new software requiring the information about CVC site (femoral or jugular/subclavian) and correcting GEDVI in case of femoral injection. Although jugular and subclavian vein access might be the preferred sites for CVC insertion, several recent studies on catheter related bloodstream infections have demonstrated that femoral venous access was used in about 20 to 35% of all catheter insertions [17, 18]. Therefore, results of TPTD might be substantially altered in about one third of cases without appropriate correction.

To the best of our knowledge there are no data available if the new PiCCO algorithm also corrects unindexed GEDV and PVPI in case of femoral CVC.

Furthermore, data on PVPI were not investigated in both previous papers on TPTD using a femoral CVC [15, 16].

Therefore, it was the aim of our study to investigate if in case of femoral indicator injection

GEDV in the new PiCCO algorithm is given based on the correction of GEDVI and

PVPI is based on corrected GEDV and corrected pulmonary blood volume.

The study was approved by the institutional review board (Ethikkommission der Fakultät für Medizin der Technischen Universität München, Munich, Germany).

All patients or their legal representatives gave written informed consent.

We prospectively performed 110 TPTDs in 11 patients with femoral indicator injection. All patients were recruited between October 2013 and January 2014.

Ten triplicate TPTDs per patient were performed with 15 mL cold saline within a total of 30 h. TPTD was performed as described previously using the PiCCO-2 device (Pulsion Medical Systems, Feldkirchen, Germany) [6, 16, 19] and a 5-lumen CVC (Multicath 5; Vygon; Aachen, Germany). All PiCCO-2 devices were equipped with the V3.1 algorithm requiring information about the venous catheter site which was set to “femoral” CVC in all patients.

Immediately after triplicate TPTD absolute and indexed mean values were documented for GEDV, GEDVI, EVLW, EVLWI and PVPI as displayed by the device. Parameters displayed by the device were subscripted with “displayed” (see Table 1).

The main purpose of the study was to investigate the consistency of displayed values by comparing displayed data (e.g. GEDV_displayed, PVPI_displayed) to the corresponding values derived from calculation specified with the subscript “calculated” (GEDV_caculated, PVPI_calculated).

If displayed and calculated values were inconsistent, we investigated whether the inconsistency could be explained by (non-) application of the recently suggested correction formula for the displayed values parameters [16]. In this case “ex post” correction by our formula should result in a consistency of the “corrected” values (with the subscript “corrected”) and the calculated values.

To analyze if the femoral TPTD correction formula for GEDVI_displayed was only applied for GEDVI_displayed or if also GEDV_displayed is corrected in case of femoral indicator injection, we calculated GEDV_calculated using the formula:

Finally, GEDV_displayed was compared to GEDV_calculated.

In a second step we analyzed, if PVPI_displayed is calculated based on a GEDV_displayed corrected for femoral indicator injection. Therefore, we compared PVPI_displayed to PVPI_calculated using the formula PVPI_calculated = EVLW_displayed/(0.25 * GEDV_displayed).

Predicted bodyweight BW_predicted was calculated using the formula:

Predicted body surface area BSA_predicted was calculated using BW_predicted instead of actual bodyweight in the Dubois formula:

The correction formula suggested for correction of uncorrected femoral indicator injection derived GEDV_uncorrected is GEDVI_corrected [mL / m²] = 0.539 * GEDVI_uncorrected - 15.17 + 24.49 * CI_uncorrected +2.311* BW_ideal[16].

Ideal bodyweight BW_ideal was calculated using the formula:

Correlation of GEDV_calculated and GEDVI_displayed was analyzed using Spearman correlation. Comparisons of GEDV_displayed vs. GEDV_calculated, PVPI_displayed vs. PVPI_calculated and PVPI_corrected vs. PVPI_calculated were performed using Wilcoxon-test for paired samples, respectively. Bland-Altman analyses and calculation of percentage error (PE) were used for these comparisons as described previously [20, 21]. Bland-Altman analyses were corrected for repeated measurements allowing variability of true values within each subject [22].

All analyses were performed using IBM SPSS Statistics 21 (SPSS inc., Chicago, IL, USA).

Despite improving outcome by a number of specific and unspecific approaches such as low tidal volume [23], extracorporeal lung assist [24] and prone positioning [25] ARDS still carries a high mortality of up to 65% [23‐26]. Well balanced fluid management providing appropriate resuscitation [27] as well as avoiding fluid overload-particularly of the lungs-is another obvious approach to improve outcome in ARDS. EVLW has been validated as a significant predictor of outcome in ARDS [3, 7‐11]. However, with regard to fluid management it is important, if increased EVLW derives from inflammatory capillary leakage of the lungs or from general fluid overload and congestive heart failure. Relating EVLW to pulmonary blood volume, PVPI has been demonstrated to discriminate inflammatory from congestive pulmonary failure [13, 14]. Therefore, combined use of EVLW and PVPI has been suggested to specify therapy in acute pulmonary failure. However, there are a number of studies suggesting pitfalls in using these parameters derived from TPTD: Due to its current indexation solely to bodyweight EVLWI might be inappropriately lowered in obese patients. Therefore, weight correction formulas [19] and indexation of EVLWI to height [19, 28] have been shown to improve usefulness of EVLWI. Furthermore, femoral central vein access for TPTD indicator injection inappropriately increases GEDV(I) [15, 16]. Therefore, a correction formula for GEDVI in case of femoral CVC has been suggested [16]. Requirement of information about the CVC site in the last PiCCO algorithm suggests that some kind of correction for femoral CVC is provided. Due to the assumption that PBV is 25% of GEDV in single indicator TPTD [6, 12], overestimation of GEDV in femoral indicator TPTD indicator injection also pertains to PVPI.

However, no information is available,

– if correction provided by the new PiCCO-algorithm applies only to GEDVI or also to GEDV

– if this correction is similar to the formula suggested by Saugel et al. [16] and

– if PVPI_displayed is based on PBV and GEDV derived from the correction formula suggested by our group.Therefore, we performed a prospective study in patients with femoral CVC access showing the following main results that are graphically summarized in Figure 6.

GEDV_displayed is corrected to the same degree as GEDVI_displayed.

However, PVPI_displayed provided by the PiCCO-device is significantly lower than PVPI_calculated derived by dividing EVLW_displayed by 0.25*GEDV_displayed. This suggests the use of uncorrected GEDV and uncorrected PBV for the PVPI displayed by the device.

This assumption is supported by PVPI_corrected values nearly identical to PVPI_calculated when correcting PVPI_displayed using the correction formula.

This study suggests that the V3.1 software of the PiCCO-device corrects GEDVI and GEDV for TPTD indicator injection based on a correction formula similar to the correction recently suggested [16]. However, correction of GEDV(I) does not pertain to PVPI which is contradictory to the definition of PVPI and results in a substantial underestimation of PVPI. To make PVPI values derived from jugular and femoral indicator injection comparable, corrected GEDV and PBV should be used for calculation of PVPI in case of femoral CVC.