Estimation of errors in determining intrathoracic blood volume using thermal dilution in pigs with acute lung injury and haemorrhage{dagger}

M. Nirmalan1,2,*, M. Niranjan3, T. Willard2,4, J. D. Edwards2, R. A. Little2 and P. M. Dark2,5

1 Critical Care Unit, Manchester Royal Infirmary, Manchester, UK. 2 MRC Trauma Group, University of Manchester, Manchester, UK. 3 Department of Computer Sciences, University of Sheffield, Sheffield, UK. 4 North Western Medical Physics Department, South Manchester University Hospitals, Manchester, UK. 5 Critical Care Unit, Hope Hospital, Manchester, UK

* Corresponding author: University Department of Anaesthesia and Critical Care Medicine, Manchester Royal Infirmary, Oxford Road, Manchester M13 9WL, UK. E-mail: m.nirmalan{at}man.ac.uk

Accepted for publication May 4, 2004.


    Abstract
 Top
 Footnotes
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
Background. Global end diastolic volume (GEDV) has a constant and predictable relationship to intrathoracic blood volume (ITBV). The present study assesses the difference between ITBV derived from GEDV and ITBV measured directly in pigs with acute lung injury (ALI) and mild haemorrhage.

Methods. We caused ALI in 12 anaesthetized pigs by i.v. injection of oleic acid and removed 10% of estimated blood volume. EVLW, GEDV, ITBV (COLD; Pulsion Medical Systems), , lung compliance and haemodynamic variables were measured at baseline (time 0) and at 30 and 120 min. All animals were volume-resuscitated, followed by measurements at 180 min. A linear equation estimated from the 44 pairs of ITBV and GEDV values in 11 animals was applied iteratively to the four GEDV measurements in the 12th animal, enabling 48 comparisons between measured (ITBVm) and derived ITBV (ITBVd) to be made.

Results. Increase in extravascular lung water index (EVLWi) was associated with significant pulmonary hypertension, worsening of oxygenation and compliance (repeated measures ANOVA; P<0.05). There was good within-subject correlation and agreement between ITBVm and ITBVd (r=0.72, mean bias 0.8 ml; SD 32 ml). Mean error in deriving ITBV from GEDV was 4.5%. (SD 4.2%; range 0.05–19%). There were no significant differences in errors in the presence of small (up to 10%) deficits in blood volume (F=1.0; P=0.41).

Conclusions. ITBV estimated by thermodilution alone is comparable to measurements made by the thermo-dye dilution technique in the presence of pulmonary hypertension and mild deficits in total blood volume.

Keywords: blood, volume ; lung, water ; measurement techniques, thermodilution ; model, pig


    Introduction
 Top
 Footnotes
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
The transpulmonary indicator dilution technique provides a useful alternative to pulmonary artery-based haemodynamic monitoring in critically ill patients. In addition to providing an estimate of cardiac preload,13 this method also allows the estimation of extravascular lung water content (EVLW), which is a more sensitive measure than chest X-ray or indices of oxygenation in detecting early pulmonary oedema.45 There is now a body of evidence which suggests that management protocols using EVLW as a therapeutic end-point may lead to better clinical outcomes in subgroups of patients who are prone to circulatory volume overload.67 A thermo-dye dilution method using temperature (T) and indocyanine green dye (IcG) dilution to derive intrathoracic thermal volume (ITTV) and intrathoracic blood volume (ITBV) respectively has been used for the accurate estimation of EVLW (EVLW=ITTV–ITBV).812 Despite the methodological limitations of some of the early validation studies, the thermo-dye dilution technique remains the only feasible method for the estimation of EVLW in critically ill patients. When using this technology (COLD; Pulsion Medical Systems Munich, Germany), ITTV is derived using cardiac output (CO) and mean transit time (MTt) for temperature (ITTV=COxMTtT) and ITBV is derived from CO and MTt for IcG (ITBV=COxMTtIcG). A transpulmonary thermodilution technique (rather than pulmonary artery thermodilution) using temperature dilution curves plotted in the abdominal aorta or common iliac artery is used to measure CO. The underlying principles of estimating EVLW using this technology have been reviewed by many authors.11213 The basic assumption is that the negative thermal energy from the cold bolus will rapidly distribute in the entire fluid compartment, whereas the IcG is protein-bound and hence will be confined to the blood volume. However, the thermo-dye dilution method, though effective, is expensive and time-consuming for routine use in a busy clinical setting.

The above limitations led to the development of an alternative method for estimating EVLW based on transpulmonary thermodilution alone (PiCCO; Pulsion Medical Systems) by utilizing the linear relationship between global end diastolic volume (GEDV) and ITBV demonstrated in previous studies in which the thermo-dye dilution method was used.89 The central assumption in this approach is that GEDV (total volume of blood in all cardiac chambers) has a constant and predictable relationship to ITBV.9 Since GEDV may be derived using the MTt and the exponential down-slope time of the thermodilution curve alone, the need for IcG dilution is eliminated. ITBV includes both GEDV and pulmonary blood volume (PBV), and therefore the above assumption implies that the relationship between GEDV and ITBV (GEDV+PBV) is constant even in the presence of all physiological derangements seen in critical illness. There is some evidence that this assumption may be flawed in the presence of regional pulmonary hypoperfusion.14 Furthermore, validation studies used9 to develop the model for the derivation of ITBV did not define the severity of lung injury, the state of the pulmonary circulation or the volume status of patients at the time of measurement. In this context it is relevant that one of the early clinical studies that included hypovolaemic patients concluded that ‘significant discrepancies exist between the two techniques’ (double- and single-indicator dilution techniques) ‘and we cannot recommend the use of single indicator dilution technique to estimate EVLW’.10

Pulmonary hypertension in acute lung injury could potentially affect the relationship between GEDV and ITBV. Unrecognized small deficits in total blood volume (seen frequently in such patients) may further contribute to redistribution of blood from pulmonary vascular beds to more central compartments and such compensatory vasoconstriction could also alter the quantitative relationship between ITBV and GEDV. We believe that if single thermodilution technology is to receive wider application in haemodynamic monitoring it is necessary to define the extent of errors in the estimation of ITBV in the presence of the physiological derangements frequently seen in critically ill patients. The present study was therefore undertaken to evaluate the errors inherent in estimating ITBV using the transpulmonary thermodilution technique in a laboratory model of critical illness associated with acute lung injury (ALI) and mild deficits in total blood volume. We compared ITBV derived from GEDV with ITBV measured directly from IcG dilution in pigs with acute lung injury and mild haemorrhage. The temperature and IcG dilution curves used in the analyses were obtained using the thermo-dye dilution technique (COLD).


    Methods
 Top
 Footnotes
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
After University Ethics Committee approval, the study was carried out in 12 immature female Large White pigs (mean weight 25.8 kg, SD 2.5 kg). Anaesthesia was induced with halothane, oxygen and nitrous oxide administered via a snout mask. After tracheal intubation mechanical ventilation was established with a volume-cycled ventilator (Blease-Brompton-Manley, Chesham, UK; tidal volume 10–15 ml kg–1; rate 12–15 b.p.m.). Anaesthesia was maintained using an i.v. infusion of alphaxalone–alphadolone (Saffan; Pitman-Moore, Uxbridge, UK; 15 mg kg–1 h–1) and all animals received maintenance fluids (0.9% NaCl; 10 ml kg–1 h–1). A pulmonary artery floatation catheter (Baxter Swan-Ganz CCO/VIP, 7.5F; Edwards Life Sciences, CA, USA) was sited via the right external jugular vein and the dye dilution catheter (Pulsiocath PV 2024 4F; Pulsion Medical Systems) was positioned in the abdominal aorta via the femoral route. An oesophageal balloon was placed in the mid-oesophagus and a suprapubic cystostomy was carried out. At the end of the instrumentation phase all animals were given a rest period of 30 min, after which baseline measurements were made (time 0).

All animals then received an infusion of oleic acid (OA; 0.05 ml kg–1) directly into the right atrium followed by the removal of 10% blood volume. The total blood volume was estimated to be 75 ml kg–1 in these animals based on previous studies in our laboratory,15 and 10% of this estimated volume was removed in each of the animals using a roller pump. A second set of haemodynamic and EVLW measurements was made after ALI and haemorrhage were established (measurements at 30 min) and the maintenance fluid infusion was stopped. The animals were allowed to remain undisturbed for a further period of 90 min followed by the third set of haemodynamic/EVLW measurements (measurements at 120 min). The animals were then volume-resuscitated over a period of 60 min followed by a final set of haemodynamic and EVLW measurements (measurements at 180 min). The end-points for resuscitation were to achieve and maintain cardiac output above the corresponding baseline values during the resuscitation phase. The experiment was terminated at this stage and the animals were killed by anaesthetic overdose and lungs removed for gravimetric estimation of lung water using wet–dry lung weights. The thermo-dye dilution method (COLD Z-03; Pulsion Medical Systems) using duplicate injections of 10 ml cold indocyanine green was used to estimate EVLW at times 0, 30 min, 120 min and 180 min respectively. If the differences in EVLWi between the duplicate injections was >10% a third injection was made, in keeping with current clinical practice, and the two closest values for EVLWi was used in all subsequent calculations. All injections were standardized to the end-expiratory phase of the ventilatory cycle. At each of the above time points other parameters of haemodynamic status and lung function, including cardiac output (using the pulmonary artery thermodilution technique), mean pulmonary arterial pressure (MPAP), pulmonary vascular resistance (PVR), and a proxy measure of lung compliance (Cpeak pressure) based on peak and end-expiratory transpulmonary pressures, were measured.

An endotracheal tube with a purpose-built pressure monitoring port was used to intubate the trachea (Mallinckrodt Hi-Lo Jet; size 6.5). An oesophageal balloon (Mallinckrodt, size 8.5) was positioned in the mid-oesophagus and the pressure monitoring ports of both tubes were connected to a differential pressure transducer (Pneu-01; World Precision Instruments, USA) to obtain transpulmonary pressure (TPP) signals. A pneumotachograph (Godart 17212; Gould Electronics, Netherlands) was used for respiratory flow measurements. Transpulmonary pressure and respiratory flow signals were acquired and stored in a personal computer using standard equipment and software (CED 1902, CED 1401 and Spike 2; Cambridge Electronics Design, Cambridge, UK). Tidal volume (Vt) was obtained by integrating the inspiratory flow signals and Cpeak pressure was calculated using the equation [Cpeak pressure=Vt/(peak TPP–end-expiratory TPP)], used in similar studies by our group previously.16 Post-mortem examination was carried out in all animals to confirm the presence of ALI.

Statistical analyses and error calculations
All variables were analysed using analysis of variance (ANOVA) for repeated measurements (general linear model; SPSS 9.0, SPSS, Chicago, IL, USA). Significant factors were compared further using the 95% confidence intervals (CI) and statistical significance was defined as P<0.05 (two-sided). Within-subject standard deviation (Sw) of duplicate measurements of EVLWi was used as a measure of reproducibility.17 A linear regression equation was estimated using 44 paired measurements of ITBV and GEDV obtained in 11 animals (four sets of values for each animal) and this model was applied to measured GEDV values (at the four time points) of the 12th animal to derive the corresponding derived ITBV (ITBVd). The process was repeated in each of the 12 animals, allowing a total of 48 comparisons between measured and derived ITBV (ITBVm and ITBVd respectively) to be made. The regression equation developed using this technique is not influenced by data from the animal in which the equation will be applied for deriving ITBVd. This out-of-sample prediction technique is in the spirit of the ‘leave one out’ cross-validation technique.18 ITBVm and ITBVd were compared using within-subject correlation(r) and the Bland–Altman plot.19 All data were normally distributed; hence, the mean (SD) and the corresponding 95% CI were used as summary statistics.


    Results
 Top
 Footnotes
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
The lungs showed oedema, diffuse erythema and multiple surface haemorrhage of varying degrees in all animals on post-mortem examination. Histological examination was undertaken in one of the animals and showed pulmonary oedema, interstitial oedema, haemorrhage and microthrombi in pulmonary vessels.20 Gravimetrically determined lung water data were available in only eight animals and ranged between 8.8 and 17.3 ml kg–1 [mean 11.6 (SD 2.7) ml kg–1]. In these eight animals, EVLW detected by the thermal-dye dilution method was on average 82% (SD 7%) of lung water detected by the gravimetric method. Administration of OA resulted in an immediate increase in MPAP and PVR associated with a significant increase in EVLWi, and worsening of oxygenation and pulmonary compliance. The relevant haemodynamic and pulmonary function data are shown in Table 1. Reproducibility of EVLWi measurements was within clinically acceptable limits [mean Sw 6.2% (SD 6.5%)]. For the pooled data from all 12 animals there was a good correlation between measured GEDV and ITBV values (ITBV=1.1*GEDV+99.5; P<0.0001; r2=0.82). In view of this correlation (r2=0.82) a simple linear transformation was considered adequate to derive ITBVd from GEDV values for each of the 12 animals. The slope and intercept for the 12 regression equations were very similar (slope range, 1.1–1.0; intercept range, 99.3–100.4). There was a significant within-subject correlation between ITBVm and ITBVd (r=0.72, P<0.001). Bias and limits of agreement between ITBVm and ITBVd derived from the Bland–Altman plot were 0.8 ml and 66.2 to –64.6 ml respectively. There were no significant differences between the percentage errors at the different stages of the experiment with changes in total circulatory volume (F=1.0; P=0.41). Overall, the errors inherent in estimating ITBV from measured GEDV were small (mean error 4.5%; SD 4.2%; range 0.05–19%). The distribution of errors in estimating ITBV from GEDV at each of the time points is shown in Fig. 1.


View this table:
[in this window]
[in a new window]
 
Table 1 Mean (SD) and 95% CI of haemodynamic, pulmonary function data and percentage errors when ITBV is derived from GEDV at baseline (time 0), 30 min, 120 min and 180 min.

 


View larger version (16K):
[in this window]
[in a new window]
 
Fig 1 Distribution of percentage error in deriving ITBV from measured GEDV values (mean 4.5%, SD 4.2%, range 0.05–18.9%). The 95% confidence intervals for the above errors were 1.9–4.7%, 2.0–10.2%, 1.9–6.7% and 2.1–6.4% at 0, 30, 120 and 180 min respectively. Even though the figure suggests a possible increase in prediction errors at 30 min, the differences were not statistically significant (Table 1).

 

    Discussion
 Top
 Footnotes
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
This study confirms that an accurate estimate of ITBV may be obtained through fixed transformation of GEDV. This approach, used in the single thermodilution technique (PiCCO), has not been validated in the presence of severe pulmonary hypertension associated with ALI and mild deficits in total blood volume. These conditions frequently coexist in many patients and pulmonary vasoconstriction seen under such conditions may potentially alter the relationship between GEDV and ITBV (GEDV+PBV). It is therefore important to quantify the errors inherent in this process. Our data show that errors were minimal (mean error 4.5%; SD 4.2%) and within clinically tolerable limits, and were not altered significantly by small changes in total blood volume (F=1.0, P>0.05). These errors will be reflected by errors of a similar magnitude in the final EVLW measurement as EVLW is the difference between ITTV and ITBV. In the clinical context, however, estimation of EVLW using thermodilution alone involves many additional stages, including determining ITTV using the MTtT, estimation of GEDV by the use of the down-slope portion of the thermodilution curve and measurement of CO using transpulmonary (rather than pulmonary arterial) thermodilution. Each of the above additional steps will invariably be associated with measurement errors, and these errors were not addressed in the present study. The present analysis, therefore, though important, is limited in scope. The cumulative effects of multiple measurement errors should be taken into account in clinical applications of single thermodilution technology for the estimation of EVLW.

The choice of anaesthetic technique has been discussed previously.16 Gravimetric methods obviously provide the gold standard measure of lung water. The most accurate gravimetric method, as described by Pearce and colleagues,21 is not valid in the presence of significant pulmonary haemorrhage and hence was inappropriate for the present study. The wet–dry lung weight method unfortunately does not distinguish between EVLW and pulmonary blood volume and hence would in effect overestimate EVLW. Therefore, the percentage of true EVLW detected by the dye-thermodilution technique is likely to be greater than reported in the present study (mean 82%, SD 7%). The severity of lung injury caused by oleic acid is dose-dependent22 and the dose of OA used in the present study was, by design, small compared with other similar studies23 in order to minimize the severity of lung injury. This is reflected in the modest, albeit significant, increase in EVLWi. Since greater blood loss is unlikely to be tolerated in the presence of an OA-induced lung injury, it was decided to restrict the haemorrhage to 10% of circulatory volume. This degree of blood loss, as expected, did not lead to significant changes in cardiac output, stroke volume, ITBV, GEDV or PBV. Even though a small reduction in cardiac output was evident at 30 and 120 min, the differences were small compared with the corresponding baseline values. Table 1 confirms that the animals achieved effective compensation for the relatively small loss in blood volume, with GEDV, ITBV and SV returning to the baseline values at 120 min, even before fluid resuscitation. Mean PBV, on the other hand, remained low at 120 min and returned to baseline values only after volume resuscitation. The absence of significant changes in cardiac output and ITBV is a potential limitation of the present study and may account for the predictable relationship between GEDV and ITBV at 30 and 120 min in spite of significant pulmonary hypertension. In the clinical context, however, larger deficits in circulatory volume or cardiac output are likely to be recognized and corrected as an immediate priority using GEDV (or other measurements, such as central venous pressure or arterial pressure wave oscillations) values before EVLW measurements become relevant. Therefore, the errors introduced by unrecognized mild hypovolaemia and pulmonary hypertension are arguably of greater concern in the clinical application of the single thermodilution technique than similar errors in the presence of gross hypovolaemia. We therefore did not set out to estimate the errors inherent in using this method in subjects with larger deficits in circulatory volume in the present study.

The correlation coefficient between ITBVm and ITBVd (r=0.72) is less than the correlation coefficient reported in two previous studies (0.87 and 0.96 respectively).8 9 Because repeated measurements were undertaken, it was important to determine whether a change in ITBVm was reflected in a corresponding change in ITBVd. We therefore used the within-subject correlation coefficient (r) in order to provide an estimate of the size and direction of changes within each subject. This measure is conceptually different24 from the simple correlation coefficient reported in the two previous studies.8 9 In clinical monitoring, within-subject changes are usually more relevant than absolute values in different subjects, and this is the first study in which within-subject changes in ITBVm and ITBVd have been compared and prediction errors quantified. Lewis and colleagues12 and Wickerts and colleagues25 showed that EVLW measurements determined by the thermo-dye dilution technique were independent of cardiac output and functioned well over the entire range of cardiac output and pulmonary oedema seen in clinical practice.12 Nevertheless, the potential dependence on homogeneous pulmonary perfusion has been cited as an impediment to the widespread acceptance of this technology.26 In fact, some authors have indicated that this method should be validated for the different subgroups of patients in the ICU.26 The PiCCO system, by its reliance on an indirect estimate of ITBV, raises further doubt about the accuracy of EVLW measurements. These issues have not been addressed satisfactorily in the current literature and were the primary objectives of our study. In a similar study, Neumann found a good correlation between measured and derived ITBV.8 This study was carried out in normovolaemic pigs with a single measurement of EVLW before and after lung injury. The present study differs from the Neumann study in many important aspects. First, repeated measurements were undertaken as lung injury evolved and all animals had a 10% deficit in total blood volume at two of the four time points. Furthermore, by applying an out-of-sample prediction technique the present analysis separates the source of the predictive model from the subject in which the actual measurement was made. Our findings are therefore more relevant to the way ITBV and EVLW measurements are made in clinical practice using the single thermodilution technique. The relationship between EVLW and indices of oxygenation has been studied by Hachenberg and colleagues,27 who found no correlation between EVLW and any of the indices of oxygenation. In the present study, however, an increase in EVLWi was associated with an acute deterioration in oxygenation and pulmonary compliance.

In conclusion, this study shows that, even in the face of pulmonary hypertension and small deficits in total blood volume, ITBV may be estimated with reasonable certainty by the single thermodilution technique. The errors inherent in estimating ITBV (and therefore EVLW) in the presence of an ALI and associated pulmonary hypertension were small. These conclusions, however, cannot be extended to situations in which a greater deficit in circulatory volume or regional perfusion defects may be present. We suggest that further experimental studies are necessary to evaluate prediction errors in the presence of circulatory shock associated with major changes in ITBV and GEDV.


    Acknowledgments
 
We thank T. Riney and H. Marshall of the MRC Trauma Group for their help in conducting these experiments, Professor A. J. Freemont, University of Manchester, for carrying out the microscopic examinations, and MRC (UK) and Maelor Pharmaceuticals Ltd for funding these experiments.


    Footnotes
 Top
 Footnotes
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
{dagger} Presented in part at the Anaesthetic Research Society Meeting in Manchester, UK, November 20, 2003. Back


    References
 Top
 Footnotes
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
1 Goedje O, Peyerl M, Seebauer T, Dewald O, Reichart B. Reproducibility of double indicator dilution measurements of intrathoracic blood volume compartments, extravascular lung water and liver functions. Chest 1998; 113: 1070–7[Abstract/Free Full Text]

2 Lichtwarck-Aschoff M, Zeravik J, Pfeiffer UJ. Intrathoracic blood volume accurately reflects circulatory volume status in critically ill patients with mechanical ventilation. Intensive Care Med 1992; 18: 142–7[ISI][Medline]

3 Nirmalan M, Willard T, Walsh K, Columb MO, Edwards JD, Little RA. Haemodynamic monitoring in a combined trauma-hypovolaemic shock model. Br J Anaesth 2001; 87: 173P–174P

4 Halperin BD, Feeley TW, Mihm FG, Chiles C, Guthaner DF, Blank NE. Evaluation of the portable chest roentgenogram for quantifying extravascular lung water in critically ill adults. Chest 1985; 88: 649–52[Abstract]

5 Baudendistel L, Shields JB, Kaminski DL. Comparison of double indicator thermodilution measurements of extravascular lung water (EVLW) with radiographic estimation of lung water in trauma patients. J Trauma 1982; 22: 983–8[ISI][Medline]

6 Mitchell JP, Schuller D, Calandrino FS, Schuster DP. Improved outcome based on fluid management in critically ill patients requiring pulmonary artery catheterisation. Am Rev Respir Dis 1992; 145: 990–8[ISI][Medline]

7 Eisenberg PR, Hansbrough JR, Anderson D, Schuster DP. A prospective study of lung water measurement during patient management in an intensive care unit. Am Rev Respir Dis 1987; 136: 662–8[ISI][Medline]

8 Neumann P. Extravascular lung water and intrathoracic blood volume: double versus single indicator dilution technique. Intensive Care Med 1999; 25: 216–9[CrossRef][ISI][Medline]

9 Sakka SG, Ruhl CC, Pfeiffer UJ, et al. Assessment of cardiac preload and extravascular lung water by single transpulmonary thermodilution. Intensive Care Med 2000; 26: 180–7[CrossRef][ISI][Medline]

10 Schuster DP, Calandrino FS. Single versus double indicator dilution measurements of extravascular lung water. Crit Care Med 1991; 19: 84–8[ISI][Medline]

11 Mihm FG, Feeley TW, Rosenthal MH, Lewis F. Measurement of extravascular lung water in dogs using the thermal-green dye indicator dilution method. Anaesthesiology 1982; 57: 116–22[ISI][Medline]

12 Lewis FR, Elings VB, Hill SL, Christensen JM. The measurement of extravascular lung water by thermal-green dye indicator dilution. Ann N Y Acad Sci 1982; 384: 394–410[Abstract]

13 Goedje O, Seebauer T, Peyerl M, Pfeiffer UJ, Reichart B. Haemodynamic monitoring by double-indicator dilution technique in patients after orthotopic heart transplantation. Chest 2000; 118: 775–81[Abstract/Free Full Text]

14 Schreiber T, Huter L, Schwarzkopf K, et al. Lung perfusion affects preload assessment and lung water calculation with the transpulmonary double indicator method. Intensive Care Med 2001; 27: 1814–8[CrossRef][ISI][Medline]

15 Foex BA, Quinn JV, Little RA, Shelly MP, Slotman GJ. Differences in eicosanoid and cytokine production between injury/haemorrhage and bacteraemic shock in the pig. Shock 1997; 8: 276–83[ISI][Medline]

16 Nirmalan M, Willard T, Edwards JD, Dark P, Little RA. Effects of sustained post-traumatic shock and initial fluid resuscitation on extravascular lung water content and pulmonary vascular pressures in a porcine model of shock. Br J Anaesth 2003; 91: 224–32[Abstract/Free Full Text]

17 Bland JM, Altman DG. Measurement error. BMJ 1996; 313: 744[Free Full Text]

18 Bishop CM. Neural Networks for Pattern Recognition. Oxford University Press, 1995; 372–5

19 Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; 1: 307–10[ISI][Medline]

20 Nirmalan M. Quantifying lung injury in critical illness. PhD thesis, University of Manchester, 2001.

21 Pearce ML, Yamashita J, Beazell J. Measurement of pulmonary oedema. Circulation Research 1965; 16: 482–8[ISI]

22 Hinz J, Neumann P, Dudykevych T, et al. Regional ventilation by electrical impedance tomography. Chest 2003; 124: 314–22[Abstract/Free Full Text]

23 Perchiazzi G, Hogman M, Rylander C, Giuliani R, Fiore T, Hedenstierna G. Assessment of respiratory system mechanics by artificial neural networks: an exploratory study. J Appl Physiol 2001; 90: 1817–24[Abstract/Free Full Text]

24 Bland JM, Altman DG. Calculating correlation coefficients with repeated observations. Part 1: correlation within subjects. BMJ 1995; 310: 446[Free Full Text]

25 Wickerts CJ, Jakobsson J, Frostell C, Hedenstierna G. Measurement of extravascular lung water by thermal-dye dilution technique: mechanisms of cardiac output dependence. Intensive Care Med 1990; 16: 115–20[ISI][Medline]

26 Boeck JC, Lewis FR. Clinical relevance of lung water measurement with the thermal-dye dilution technique. J Surg Res 1990; 48: 254–65[CrossRef][ISI][Medline]

27 Hachenberg T, Tenling A, Rothen HU, Nystrom SO, Tyden H, Hedenstierna G. Thoracic intravascular and extravascular fluid volumes in cardiac surgical patients. Anesthesiology 1993; 79: 976–84[ISI][Medline]





This Article
Abstract
Full Text (PDF)
All Versions of this Article:
93/4/546    most recent
aeh232v1
E-Letters: Submit a response to the article
Alert me when this article is cited
Alert me when E-letters are posted
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Search for citing articles in:
ISI Web of Science (1)
Disclaimer
Request Permissions
Google Scholar
Articles by Nirmalan, M.
Articles by Dark, P. M.
PubMed
PubMed Citation
Articles by Nirmalan, M.
Articles by Dark, P. M.