Intensive Care Medicine

, Volume 34, Issue 11, pp 2106–2111 | Cite as

Extravascular lung water volume measurement by a novel lithium-thermal indicator dilution method: comparison of three techniques to post-mortem gravimetry

  • Benjamin Maddison
  • Riccardo Giudici
  • Enrico Calzia
  • Christopher Wolff
  • Charles Hinds
  • Peter Radermacher
  • Rupert M. Pearse
Brief Report



To compare the lithium-thermal double indicator dilution (Li-thermal), indocyanine green-thermal double indicator dilution (ICG-thermal), single thermal indicator dilution (single-thermal) and gravimetric techniques of extravascular lung water volume (EVLW) measurement in porcine models of acute lung injury.


Two animal models designed to invoke a systemic inflammatory response.


Laboratory study.


A total of 12 immature Deutsches Landschwein pigs.


Extravascular lung water volume was measured at four time points using Li-thermal, ICG-thermal and single-thermal techniques. Measurements were performed using existing technology according to manufacturer’s instructions. Post-mortem gravimetric EVLW measurements were performed by measuring wet and dry mass of lung tissue. Measurements were compared using the Bland–Altman method. Data are presented as mean (SD).

Measurements and main results

Data were collected in 12 animals and comparison between all 4 techniques was possible in 10 animals. EVLW measured by gravimetry was 9.2 (±3.0)ml kg−1. When compared to gravimetry, both Li-thermal and ICG-thermal techniques showed minimal bias but wide limits of agreement (LOA) [Li-thermal: bias −1.8 ml kg−1 (LOA ± 13.1); ICG-thermal bias −1.0 ml kg−1 (LOA ± 6.6)]. Comparison between the single-thermal and gravimetric methods identified both considerable bias and wide LOA [+8.5 ml kg−1 (LOA ± 14.5)].


Clinically significant differences between EVLW measurements obtained with the gravimetric method and three in vivo indicator dilution techniques were identified. While none of the techniques could be considered ideal, the ICG-thermal method appeared more reliable than either the Li-thermal or single thermal techniques. Further research is required to determine whether the accuracy of the prototype Li-thermal technique can be improved.


Cardiovascular monitoring Intrathoracic blood volume Extravascular lung water Lithium indicator dilution Gravimetry Acute lung injury 


Expansion of the extravascular lung water (EVLW) volume during critical illness is associated with increased morbidity and mortality [1]. Improved clinical outcomes may be achieved if therapy is adjusted in response to measurements of EVLW volume [2]. Clinical assessment and chest radiography appear to be unreliable guides to EVLW volume [3]. While a number of techniques have been developed to measure EVLW volume at the bedside, none of the available methods can be considered ideal.

The standard clinical method of EVLW volume measurement is trans-pulmonary double indicator dilution. Intrathoracic blood volume (ITBV) and intrathoracic water volume are calculated according to Stewart’s principle (Eq. 1) allowing EVLW volume to be calculated by subtraction [4, 5]. Traditionally, indocyanine green (ICG) has been used to measure ITBV while a cold saline indicator is used to determine intrathoracic thermal volume (ITTV). This method has been validated in animal models by comparison with post-mortem measurement of the wet and dry mass of the lungs (gravimetry) [6]. Unfortunately, the ICG-thermal technique has been withdrawn because it proved too expensive and cumbersome for routine use. The only commercially available clinical method of EVLW volume measurement is the trans-pulmonary single thermal indicator dilution technique (single-thermal) [7]. However, the mathematical assumptions that underpin the derivation of ITBV limit the accuracy of this method [8]. There is therefore a need for a practical and accurate alternative to current methods of EVLW volume measurement.
$$ V = {\text{CO}} \times {\text{MTT}} .$$
The volume of distribution (V) of an indicator can be calculated from the product of the cardiac output (CO) and the mean transit time of the indicator (MTT). The volume of a compartment may be measured by selecting an indicator which distributes throughout that compartment. Intrathoracic blood volume measurement therefore requires an indicator which remains within the bloodstream while intrathoracic water volume measurement requires an indicator which distributes throughout the thoracic water compartment. In practice, thoracic water is measured using a thermal indicator (cold saline) and the term intrathoracic thermal volume is therefore used.

One approach may be to use trans-pulmonary lithium indicator dilution to measure ITBV and, in combination with thermodilution, EVLW volume. Lithium chloride satisfies many of the criteria for an ideal indicator [9], having a good safety profile [10] and small displacement volume [11]. Studies have demonstrated minimal lithium indicator loss from the vascular compartment during the measurement period i.e. one circulation time [9, 12]. It is therefore possible that ITBV measurement with lithium chloride might provide a more convenient and reliable alternative to commercially available technology. The aim of this study was to compare the ICG-thermal and single-thermal methods of EVLW volume measurement to the prototype lithium-thermal (Li-thermal) method and all of these to the gravimetric reference technique in porcine models of acute lung injury.

Materials and methods

This study was performed on two porcine models designed to invoke systemic inflammatory and associated cardiovascular and respiratory responses. These were a model of faecal peritonitis induced septic shock and an aortic balloon occlusion model of visceral organ ischaemia. The observational study reported here was incorporated into two ongoing interventional laboratory studies to minimise animal use. The study protocols, described in detail in the Supplementary material, were approved by the Animal Care Committee of the University of Ulm and the Regierungspräsidium Tübingen, Germany.

Post-mortem EVLW volume was calculated by a gravimetric method described in detail in the supplementary material. Indicator dilution measurements were made at four time points in each model. Pre-intervention measurements were made in both models, then at 2, 4 and 24 h in the sepsis model and at 1, 4 and 8 h in the visceral organ ischaemia model. The final indicator dilution measurement performed immediately before killing was used for comparison with gravimetry. ICG-thermal measurements were made using the COLD-Z system, (Pulsion Medical Systems, Munich, Germany) following central injection of 10 ml of iced 5% dextrose solution containing 25 mg of ICG according to the manufacturer’s instructions [13]. The arterial changes in temperature and ICG concentration were measured using a thermistor tipped spectrophotometric catheter (PV 2024; Pulsion Medical Systems, Munich, Germany). Lithium dilution measurements were made using the LiDCOplus system (LiDCO Ltd, Cambridge, UK) following central injection of 0.3 mmol (2 ml) of lithium chloride [14]. The arterial lithium ion concentration was measured using an external lithium ion sensor attached to the femoral arterial catheter via a 0.75 ml extension tube. Flow of arterial blood across the lithium sensor was regulated using a battery powered peristaltic pump. Time of injection was standardised through the use of a visual countdown on the monitor. Lithium cardiac output and lithium mean transit time (MTT) were used to calculate ITBV. Lithium cardiac output and COLD-Z saline thermodilution mean transit time were used to calculate lithium-thermal ITTV. Single-thermal dilution measurements of ITBV, ITTV and therefore EVLW volume were made using the PiCCO system (Pulsion Medical Systems, Munich) following central injection of 10 ml of ice cold saline according to the manufacturer’s instructions [15]. The temperature change of arterial blood was measured using a thermistor tipped catheter (PV2015L20; Pulsion Medical Systems, Munich, Germany). Cardiac output was also measured using a pulmonary artery (PA) catheter at each time-point by injecting 10 ml iced normal saline at the end of expiration (Edwards Lifesciences, Irvine, USA). Three measurements were obtained and then averaged using a cardiac output computer (Sat-2; Baxter Edwards Lifesciences, Irvine, USA). All indicator and thermodilution measurements were performed within a period of 10 min at each time point.

Statistical analysis

We calculated that, with a type I error rate of 5% and a type II error rate of 10%, 48 comparisons in 12 animals would allow the detection of a 1.0 ml kg−1 (SD ±2.0 ml kg−1) difference in EVLW volume between the Li-thermal and ICG-thermal techniques. ITBV and EVLW volume measurements were compared using the technique of Bland and Altman with comparisons presented as bias [95% limits of agreement (LOA)]. EVLW volume measurements comparing gravimetry to all three techniques were also analysed using linear regression. Percentage error was calculated according to the method reported by Critchley as the limit of agreement (±2SD) of the bias divided by the mean cardiac output from the two methods under comparison. Data are presented as mean (SD) where normally distributed and median [inter-quartile range (IQR)] where not normally distributed. Parametric data were compared using the unpaired t test and non-parametric data were compared with the Mann–Whitney test. Differences in measurements of EVLW volume, ITBV and cardiac output over time were compared using repeated measures analysis of variance (ANOVA) with Tukey’s correction. Analysis was performed using GraphPad Prism version 4 (GraphPad software, San Diego, USA). Significance was set at P < 0.05.


Data were collected in 12 animals. Comparative data were obtained for all measurement techniques in ten animals [weight 43.9 kg (±4.2)]. In one case, data were lost due to the early killing of an animal that did not develop sepsis and in the other because of damage to a monitoring cable. Data describing physiological changes consistent with acute lung injury are presented in supplementary Tables 1 to 5. Mean EVLW volume as measured by gravimetry was 9.2 ml kg−1 ± 3.0 (n = 10). Mean values of EVLW volume were 5.8 ml kg−1 (±7.8) for the Li-thermal method, 6.1 ml kg−1 (±3.3) for the ICG-thermal method and 13.0 ml kg−1 (±4.3) for the single-thermal method (n = 40). Bland–Altman comparisons between Li-thermal, ICG-thermal, single-thermal and gravimetric EVLW volume measurements are presented in Fig. 1. Linear regression analysis against the reference gravimetric measurements identified a poor correlation with each of the three indicator dilution techniques (Li-thermal: r 2 = 0.03, P = 0.65; ICG-thermal: r 2 = 0.31, P = 0.09; single-thermal: r 2 = 0.19, P = 0.20) (Supplementary Fig. 1). Comparison of the lithium-thermal measurements of EVLW volume to those taken with the ICG-thermal technique identified a minimal bias with wide LOA [bias −0.33 ml kg−1 (LOA ± 13.1); P = 0.8] (Fig. 2). Single-thermal measurements of EVLW volume were significantly greater than those taken with the ICG-thermal technique also with wide LOA [bias +7.0 ml kg−1 (LOA ± 10.5); P < 0.001] (Supplementary Fig. 2).
Fig. 1

Bland–Altman analysis of EVLW volume measures taken using lithium-thermal (Li-thermal), indocyanine green-thermal (ICG-thermal) and single-thermal versus gravimetric measurements. Only includes indicator dilution data from final measurement point (n = 10). Data presented as bias (±SD) and 95% limits of agreement (LOA). Dotted lines indicate bias and 95% LOA. EVLW sxtravascular lung water

Fig. 2

Bland–Altman analysis of Li-thermal versus ICG-thermal EVLW volume measurement. Data from all measurement points (n = 40). Data presented as bias (±SD) and 95% limits of agreement (LOA). Dotted lines indicate bias and 95% LOA. EVLW Extravascular lung water volume; Li-Thermal lithium thermal double indicator dilution; ICG-thermal indocyanine green-thermal double indicator dilution

Mean values of ITBV were 28.9 ml kg−1 (±6.2) for the ICG-thermal method, 21.7 ml kg−1 (±5.9) for the Li-thermal method and 24.6 ml kg−1 (±6.1) for the single-thermal method (n = 40). Measurements of ITBV were significantly less than those taken with the ICG-thermal technique for both the lithium-thermal [bias −7.3 ml kg−1 (LOA ± 16.4); P < 0.001] and single-thermal techniques [bias −4.5 ml kg−1 (LOA ± 10.4); P < 0.01] (Supplementary Fig. 3).

Mean cardiac output measured using the PA catheter was 4.5 l min−1 (±1.6) (n = 40). Equivalent values for the techniques under investigation were 4.8 l min−1 (±1.5) for lithium indicator dilution, 6.4 l min−1 (±2.2) for ICG indicator dilution and 5.8 l min−1 (±1.7) for the single-thermal method (n = 40). Lithium indicator dilution measurements of cardiac output were similar to those taken with the PA catheter [bias +0.2 l min−1 (LOA ± 1.6, percentage error 35%); P = 0.13] while both ICG dilution [bias +1.8 l min−1 (LOA ± 2.7, percentage error 50%); P < 0.001] and single thermal measurements [bias +1.3 l min−1 (LOA ± 1.7, percentage error 34%); P < 0.0001] were significantly greater than those made using the PA catheter.


The principal finding of this study was that, when compared to the gravimetric technique, both Li-thermal and ICG-thermal techniques showed minimal bias but wide limits of agreement. The single-thermal indicator technique, however, appeared to over-estimate EVLW volume with similar limits of agreement to the Li-thermal technique. These data suggest that double indicator dilution methods for EVLW volume measurement may be more reliable than the single-thermal technique. Further research is required to establish whether the accuracy of the prototype Li-thermal technique can be improved to provide a method which is as accurate but more convenient than the ICG-thermal method.

Interestingly lithium indicator dilution measurements of cardiac output agreed more closely with the PA catheter reference measurements than either ICG indicator dilution or single-thermal indicator dilution. The importance of cardiac output data to the accurate measurement of EVLW volume is illustrated by Eq. 1. A previous study has suggested that cardiac output measurement error may result in inaccurate determination of EVLW volume by ICG-thermal dilution [16], perhaps explaining the wide limits of agreement in the current study. However, the findings of a recent investigation suggest thermal indicator loss from oedematous lungs does not affect the accuracy of cardiac output measurement by indicator dilution [17]. The wide limits of agreement between the Li-thermal and gravimetric techniques are likely to be due to errors in the measurement of mean transit time, the most probable explanation being the distal position of the lithium ion sensitive electrode. Ideally, the electrode should be positioned within the aorta at the level of the diaphragm, ensuring a true measurement of mean transit time. In this study, however, an external lithium electrode was used and the MTT measurement adjusted on the assumption that the additional time taken for the indicator to travel from the margin of the thoracic cage to the external lithium sensor was constant. The development of an intra-arterial electrode, incorporating a thermistor, may therefore improve the accuracy of MTT measurements with this method and also allow measurements of EVLW.

Although considered by some to be the ‘gold standard’, the gravimetric method of EVLW volume measurement also has some limitations [18, 19]. Only one post-mortem measurement is possible and fluid re-absorption during the interval between death and lung removal may affect the accuracy of this method. However, the values we obtained were similar to those of previous studies that used the gravimetric technique in models of acute lung injury and sepsis [8, 20] and greater than those in sham operated animals [8]. Similarly, our pre-intervention ICG-thermal measurements of EVLW volume were consistent with those in control animals from a previous study [21]. Also in keeping with the current study, this previous work suggested that the ICG-thermal dilution technique underestimates EVLW volume when compared to gravimetry [21]. Such underestimates may be more frequent during acute lung injury [20] or the application of positive end-expiratory pressure [22]. It is not clear why none of the indicator dilution methods proved particularly reliable. In every case, measurements were carefully performed by experienced investigators according to the manufacturer’s instructions. While previous comparisons of the ICG-thermal and single-thermal methods suggest good agreement with gravimetric measurements [6, 23], other studies have demonstrated a need for adjustment of the algorithm dependant on the individual circumstances of the experiment [8, 24, 25].


We have demonstrated clinically significant differences between EVLW measurements obtained with the gravimetric method and three in vivo indicator dilution techniques. While none of the EVLW measurement techniques could be considered ideal, the ICG-thermal method appears to be more reliable than both the Li-thermal and single thermal techniques. Further development of the prototype Li-thermal method is required to establish whether this technology can provide EVLW volume measurement with similar reliability to existing double indicator techniques.



The authors wish to thank Mr. Eric Mills of LiDCO Ltd., for his advice during this study. R.M.P. formulated the hypothesis and developed the protocol with C.H. and P.R. The investigation was performed by B.M., R.G., E.C. and P.R. at Universitätsklinikum, Ulm, Germany. CW assisted in the data analysis. The manuscript was drafted by B.M., R.M.P. and C.H. All authors approved the final version.

Conflict of interest statement

This research was supported by an unrestricted educational grant from LiDCO Ltd., Cambridge, UK. R.M.P. has received speaking fees and equipment loans from Pulsion Medical Systems and a research grant from USCOM Ltd.

Supplementary material

134_2008_1207_MOESM1_ESM.doc (268 kb)
ESM1 (DOC 269 kb)
134_2008_1207_MOESM2_ESM.doc (41 kb)
ESM2 (DOC 41 kb)


  1. 1.
    Sakka SG, Klein M, Reinhart K, Meier-Hellmann A (2002) Prognostic value of extravascular lung water in critically ill patients. Chest 122:2080–2086PubMedCrossRefGoogle Scholar
  2. 2.
    Mitchell JP, Schuller D, Calandrino FS, Schuster DP (1992) Improved outcome based on fluid management in critically ill patients requiring pulmonary artery catheterization. Am Rev Respir Dis 145:990–998PubMedGoogle Scholar
  3. 3.
    Laggner A, Kleinberger G, Haller J, Lenz K, Sommer G, Druml W (1984) Bedside estimation of extravascular lung water in critically ill patients: comparison of the chest radiograph and the thermal dye technique. Intensive Care Med 10:309–313PubMedCrossRefGoogle Scholar
  4. 4.
    Chinard FP (1975) Estimation of extravascular lung water by indicator-dilution techniques. Circ Res 37:137–145PubMedGoogle Scholar
  5. 5.
    Stewart GN (1921) The pulmonary circulation time, the quantity of blood in the lungs and the output of the heart. Am J Physiol 58:20–44Google Scholar
  6. 6.
    Slutsky RA, Higgins CB (1985) In vivo validation of the thermal-green dye technique for measuring extravascular lung water. Crit Care Med 13:432–435PubMedCrossRefGoogle Scholar
  7. 7.
    Pfeiffer UJ, Lichtwarck-Aschoff M, Beale R (1994) Single thermodilution monitoring of global end-diastolic volume, intrathoracic blood volume and extravascular lung water (abstract). Clin Intensive Care 5(Suppl):38–39Google Scholar
  8. 8.
    Rossi P, Wanecek M, Rudehill A, Konrad D, Weitzberg E, Oldner A (2006) Comparison of a single indicator and gravimetric technique for estimation of extravascular lung water in endotoxemic pigs. Crit Care Med 34:1437–1443PubMedCrossRefGoogle Scholar
  9. 9.
    Band DM, Linton RA, O’Brien TK, Jonas MM, Linton NW (1997) The shape of indicator dilution curves used for cardiac output measurement in man. J Physiol 498:225–229PubMedGoogle Scholar
  10. 10.
    Jonas M, Linton RAF, O’Brien TK, Band DM, Linton NWF, Kelly F, Burden TJ, Chevalier SFA, Thompson RPH, Birch NJ, Powell JJ (2001) The pharmacokinetics of intravenous lithium chloride in patients and normal volunteers. J Trace Microprobe Tech 19:313–320CrossRefGoogle Scholar
  11. 11.
    Renner LE, Morton MJ, Sakuma GY (1993) Indicator amount, temperature, and intrinsic cardiac output affect thermodilution cardiac output accuracy and reproducibility. Crit Care Med 21:586–597PubMedCrossRefGoogle Scholar
  12. 12.
    Kurita T, Morita K, Kawasaki H, Fujii K, Kazama T, Sato S (2002) Lithium dilution cardiac output measurement in oleic acid-induced pulmonary edema. J Cardiothorac Vasc Anesth 16:334–337PubMedCrossRefGoogle Scholar
  13. 13.
    Pfeiffer UJ, Birk M, Aschenbrenner A, Blumel G (1982) The system for quantitating thermal-dye extravascular lung water. In: Prakash O (ed) Computers in critical care and pulmonary medicine. Plenum, New York, pp 123–125Google Scholar
  14. 14.
    Pearse RM, Ikram K, Barry J (2004) Equipment review: an appraisal of the LiDCO plus method of measuring cardiac output. Crit Care 8:190–195PubMedCrossRefGoogle Scholar
  15. 15.
    von Spiegel T, Wietasch G, Bursch J, Hoeft A (1996) Cardiac output determination with transpulmonary thermodilution. An alternative to pulmonary catheterization? Anaesthesist 45:1045–1050CrossRefGoogle Scholar
  16. 16.
    Hill SL, Elings VB, Lewis F (1981) Effect of cardiac output on extravascular lung water. Am Surg 47:522–528PubMedGoogle Scholar
  17. 17.
    Pohl T, Kozieras J, Sakka S (2008) Influence of extravascular lung water on transpulmonary thermodilution-derived cardiac output measurement. Intensive Care Med 34:533–537PubMedCrossRefGoogle Scholar
  18. 18.
    Fernandez-Mondejar E, Castano-Perez J, Rivera-Fernandez R, Colmenero-Ruiz M, Manzano F, Perez-Villares J, de la Chica R (2003) Quantification of lung water by transpulmonary thermodilution in normal and edematous lung. J Crit Care 18:253–258PubMedCrossRefGoogle Scholar
  19. 19.
    Pearce ML, Yamashita J, Beazell J (1965) Measurement of pulmonary edema. Circ Res 16:482–488PubMedGoogle Scholar
  20. 20.
    Rossi P, Oldner A, Wanecek M, Leksell LG, Rudehill A, Konrad D, Weitzberg E (2003) Comparison of gravimetric and a double-indicator dilution technique for assessment of extra-vascular lung water in endotoxaemia. Intensive Care Med 29:460–466PubMedCrossRefGoogle Scholar
  21. 21.
    Roch A, Michelet P, Lambert D, Delliaux S, Saby C, Perrin G, Ghez O, Bregeon F, Thomas P, Carpentier JP, Papazian L, Auffray JP (2004) Accuracy of the double indicator method for measurement of extravascular lung water depends on the type of acute lung injury. Crit Care Med 32:811–817PubMedCrossRefGoogle Scholar
  22. 22.
    Carlile PV, Lowery DD, Gray BA (1986) Effect of PEEP and type of injury on thermal-dye estimation of pulmonary edema. J Appl Physiol 60:22–31PubMedGoogle Scholar
  23. 23.
    Katzenelson R, Perel A, Berkenstadt H, Preisman S, Kogan S, Sternik L, Segal E (2004) Accuracy of transpulmonary thermodilution versus gravimetric measurement of extravascular lung water. Crit Care Med 32:1550–1554PubMedCrossRefGoogle Scholar
  24. 24.
    Kirov MY, Kuzkov VV, Kuklin VN, Waerhaug K, Bjertnaes LJ (2004) Extravascular lung water assessed by transpulmonary single thermodilution and postmortem gravimetry in sheep. Crit Care 8:R451–R458PubMedCrossRefGoogle Scholar
  25. 25.
    Neumann P (1999) Extravascular lung water and intrathoracic blood volume: double versus single indicator dilution technique. Intensive Care Med 25:216–219PubMedCrossRefGoogle Scholar

Copyright information

© Springer-Verlag 2008

Authors and Affiliations

  • Benjamin Maddison
    • 1
  • Riccardo Giudici
    • 2
  • Enrico Calzia
    • 3
  • Christopher Wolff
    • 1
  • Charles Hinds
    • 1
  • Peter Radermacher
    • 3
  • Rupert M. Pearse
    • 1
  1. 1.Intensive Care Unit, Royal London HospitalBarts and The London School of Medicine and Dentistry, Queen Mary University of LondonLondonUK
  2. 2.Institute of Anaesthesiology and Intensive Care MedicinePolo Universitario San Paolo, University of MilanMilanItaly
  3. 3.Sektion Anästhesiologische Pathophysiologie und VerfahrensentwicklungUniversitätsklinikumUlmGermany

Personalised recommendations