The present authors hypothesised that in severe acute respiratory distress syndrome (ARDS), pronation may reduce ventilator-induced overall stress (i.e. transpulmonary pressure (PL)) and strain of lung parenchyma (i.e. tidal volume (VT)/end-expiratory lung volume (EELV) ratio), which constitute major ventilator-induced lung injury determinants. The authors sought to determine whether potential pronation benefits are maintained in post-prone semirecumbent (SRPP) posture under pressure-volume curve-dependent optimisation of positive end-expiratory pressure (PEEP).
A total of 10 anesthetised/paralysed, mechanically ventilated (VT = 9.0±0.9 mL·kg−1 predicted body weight; flow = 0.91±0.04 L·s−1; PEEP = 9.4±1.3 cmH2O) patients with early/severe ARDS were studied in pre-prone semirecumbent (SRBAS), prone, and SRPP positions. Partitioned respiratory mechanics were determined during iso-flow (0.91 L·s−1) experiments (VT varied within 0.2–1.0 L), along with haemodynamics, gas exchange, and EELV.
Compared with SRBAS, pronation/SRPP resulted in reduced peak/plateau PL at VTs≥0.6 L; static lung elastance and additional lung resistance decreased and chest wall elastance (in prone position) increased; EELV increased (23–33%); VT/EELV decreased (27–33%); arterial oxygen tension/inspiratory oxygen fraction and arterial carbon dioxide tension improved (21–43/10–14%, respectively), and shunt fraction/physiological dead space decreased (21–50/20–47%, respectively).
In early/severe acute respiratory distress syndrome, pronation under positive end-expiratory pressure optimisation may reduce ventilator-induced lung injury risk. Pronation benefits may be maintained in post-prone semirecumbent position.
- Acute respiratory distress syndrome
- gas exchange
- lung recruitment
- mechanical ventilation
- mechanics of the respiratory system
- prone position
In mechanically ventilated patients, lung stress and strain are major determinants of ventilator-induced lung injury (VILI) 1–5. Alveolar stress (i.e. transmural pressure) is the ratio of alveolar wall tension to thickness. Thus, plateau/peak transpulmonary pressure (PL) reflects overall lung parenchymal stress 2, 3. Strain is distending force-induced lung parenchyma deformation, and corresponds to the tidal volume (VT)/end-expiratory lung volume (EELV) ratio 3.
Acute respiratory distress syndrome (ARDS) patients are VILI susceptible, due to disorder-induced lung regional collapse/consolidation, which severely constrains normally aerated lung parenchyma 6–8. VILI prevention comprises low VT use, positive end-expiratory pressure (PEEP), and prone positioning 3. Low VT ventilation (6 mL·kg−1) may not be advantageous relative to “standard” VT ventilation (8–9 mL·kg−1) 9. PEEP may cause circulatory depression 10, increase pulmonary oedema 11, 12, and contribute to VILI by inducing lung regional overdistension 12, 13. In severe ARDS with diffuse and bilateral aeration loss 8, 13, 14, PEEP-induced overinflation risk is reduced, even at high PEEP (17 cmH2O) 15. However, lung overdistension (excessive pressure applied to acutely-injured parenchyma) can occur without concomitant overinflation 14; consequently, the rationale for optimal PEEP level selection (PEEP optimising arterial oxygenation and minimising oxygen toxicity/VILI risks) still holds. Prone positioning may attenuate VILI 15–17. Prone position causes more homogenous lung inflation and eliminates lung compression by the heart and abdominal contents, thus limiting atelectasis 14, 18, 19. Semirecumbent positioning may also attenuate cardiac/abdominal lung compression relative to a supine position 14. Consequently, in severe ARDS, the elucidation of the effects of PEEP/body positioning optimisation 14 with respect to major VILI determinants might be of considerable clinical importance.
The present authors tested the hypothesis that following PEEP optimisation, prone positioning may reduce PL and VT/EELV relative to pre-prone semirecumbent (SRBAS) in patients with early/severe ARDS. The current authors also sought to determine whether the potential pronation benefits are maintained in the post-prone semirecumbent (SRPP) posture. Total respiratory system (RS), chest wall and lung mechanics, and haemodynamics/gas exchange were also determined and compared among the aforementioned body postures.
MATERIAL AND METHODS
Institutional Review Board (Evaggelismos General Hospital, Scientific Committee, Athens, Greece) approval and informed/written next-of-kin consent were obtained. A total of 10 consecutive patients with early (disorder diagnosis established within the preceding 72 h) and severe ARDS (table 1⇓) 20 were enrolled between July 14, 2003 and July 12, 2004. ARDS was defined as acute onset, arterial oxygen tension (Pa,O2)/inspiratory oxygen fraction (FI,O2) <200 mmHg (regardless of PEEP level), bilateral infiltrates on frontal chest radiograph, and wedge pressure <18 mmHg 21. Inclusion criteria with respect to ARDS severity were Pa,O2/Fi,O2≤100 mmHg, and a “white lungs” feature (hyperattenuation areas equally disseminated within upper and lower lung lobes) 20 on frontal chest radiograph. Radiographs were independently evaluated by two independent radiologists. Exclusion criteria included age <18 yrs, pregnancy, intracranial hypertension, burns >30% of body surface area, spine fractures, smoking and/or chronic respiratory disease, chronic liver disease (Child-Pugh class C) 22, neuromuscular disease impairing spontaneous breathing, sickle cell disease, and body mass index >27.5 kg·m−2.
During the 6-h duration study-period, patient care was provided by an independent physician. New/additional administration of i.v. fluid boluses, inotropes, antipyretics, antiarrhythmic treatment, or diuretics led to patient exclusion 23. Electrocardiographic lead II, peripheral intra-arterial and pulmonary artery (continuous cardiac output/Sv,O2 catheter; Baxter, Deerfield, IL, USA) pressures, urinary bladder temperature, and peripheral oxygen saturation (Sa,O2) were monitored continuously 23.
Prior to study enrolment, patients were sedated (propofol/fentanyl infused at 3.5–4.0 mg·kg−1·h−1/1.5–2.0 μg·kg−1·h−1, respectively), orotracheally intubated and mechanically ventilated (Siemens 300C ventilator; Siemens AG, Berlin, Germany) in a “near-supine” position (20–30° inclination relative to horizontal). Intermittent neuromusculuar blockade was employed according to recent recommendations 24. PEEP (13.4±1.8 cm H2O), FI,O2 (0.79±0.07), and breathing rate·min−1 (range 19–28) were set according to the ARDS Network protocol 25 (table 1⇓). VT (0.49±0.02 L = 7.4±0.9 mL·kg−1 predicted body weight (PBW)) was adjusted so that plateau tracheal pressure (P2,aw) was <35 cmH2O, or kept within 6.0–6.5 mL·kg−1 PBW if P2,aw exceeded 35 cmH2O (table 1⇓). Sa,O2 achieved was 93.4±1.1%. Following study enrolment, anaesthesia and neuromuscular blockade were induced/maintained with additional propofol/fentanyl (induction bolus = 0.5 mg·kg−1/50 μg and maintenance infusion = 3.5–4.0 mg·kg−1·h−1/1.5–2.0 μg·kg−1·h−1, respectively) and cisatracurium (intermittent administration targeted to full train-of-four inhibition throughout the study-period 23), respectively, and body posture was changed to “steep” semirecumbent (60° inclination, SRBAS). Prior to cisatracurium administration, oesophageal and gastric balloons were inserted and their correct placement was verified as previously described 26–28. Study-period baseline ventilator settings (volume-controlled mode) were: FI,O2 = 0.79±0.07 (as above); VT = 0.6±0.03 L (9.0±0.9 mL·kg−1 PBW); inspiratory flow (V′) = 0.91±0.04 L·s−1; breathing rate·min−1 = 16–25 (adjusted to maintain an arterial pH>7.30) 24; plateau pressure time = 0 s; and PEEP = 9.4±1.3 cm H2O. PEEP was set at 2.0 cmH2O above the lower inflection points of the pressure-volume curves, which were constructed as previously reported 20. PEEP adjustment should not cause a Pa,O2/FI,O2 decrease of >5 mmHg. Employed PEEP/VT should not result in a plateau PL (P2,L) >30 cmH2O. SRPPSa,O2 was 92.0±1.0%.
Protocol and measurements
Investigational interventions were separated by 15 min of baseline ventilation for the re-establishment of baseline conditions 29. Patients were sequentially studied in the SRBAS, prone (0° inclination), and SRPP (60° inclination) postures (2-h study duration for each posture). Patient turning was performed as previously described 23. Following pronation, abdominal movement restriction was minimised as in previous studies 23, 30. Any pronation-induced hypoxaemia (Sa,O2≤90%) would result in protocol termination, and body posture change with ventilatory parameter adjustment as necessary. The reliability of oesophageal pressure (Poes) measurements was tested as before (fig. 1⇓) 23.
The Poes tracing in figure 1d⇓ is the average of the tracings of the four sigh test breaths administered to the patient; all other tracings originate from the first of the four test sighs. On individual test breath Poes tracings (figure 1e⇓), the maximal amplitude of each cardiac oscillation was measured as the difference between oscillation peak pressure (Ppeak) and the immediately preceding inspiratory plateau pressure (Pi,plateau). For each cardiac oscillation, Poes rise rate was determined as the time needed for Poes to rise from Pi,plateau to Ppeak during that particular oscillation. For each Poes tracing, the mean maximal amplitudes of all cardiac oscillations, and mean Poes rise rates during these oscillations were analysed; for each set of test breaths, the aforementioned variables were averaged and compared among study postures.
Inspiratory flow (V′), VT, and tracheal (Paw), Poes, and gastric (Pga) pressures were measured with a Hans-Rudolph pneumotachograph (pneumotachometer; Hans Rudolph Inc., Kansas City, MT, USA) and Validyne pressure-transducers (Validyne, Nortridge, CA, USA) 23. Following analogue-to-digital conversion, variable data were stored on an IBM-type computer for later analysis with Anadat software (RHT-InfoData, Montreal, QC, Canada) 23. During data sampling, variable tracings were displayed on a dedicated monitor and recorded (Gould ES 1000 electrostatic recorder; Gould Electronics Inc., Eastlake, OH, USA). Breathing circuit modifications included humidifier-removal and low compliance tubing 23. Care was taken to avoid gas leaks. Equipment dead space (endotracheal tube (ETT) not included) was 90 mL.
Respiratory mechanics were assessed with constant V′ rapid airway occlusion. Within 30–60 min after study-posture assumption, sets of four test breaths were administered with a constant, square-wave V′ (0.91 L·s−1). VT was sequentially varied from 0.6 (baseline) to 0.2, 0.4, 0.6, 0.8, and 1.0 (“sigh”) L. During sighs, the maximum allowable P2,L was 45 cm H2O. Test breaths were separated by 1-min baseline ventilation periods (fig. 1⇓).
Test breaths were preceded by 2-s duration end-expiratory occlusions, enabling determination of RS, chest wall, and abdominal chest wall-component total PEEP (PEEPtot) (fig. 1⇓); the latter was always ∼0. For Poes, end-expiratory occlusion-plateaus were obtained by ensemble averaging 31 of Poes tracings of each test breath-set (fig. 1⇓). Expiratory occlusions were followed by 4–6-s duration end-inspiratory occlusions, enabling determination of maximal pressure (Pmax), and pressure immediately after initiation of end-inspiratory occlusion (P1), and plateau pressure (P2) on computer-stored Paw tracings, and of Pmax and P2 on computer-stored Pga tracings (fig. 1⇓). For Poes, Pmax/P1 and P2, values were determined after ensemble tracing-averaging of each set of test breaths 31 (fig. 1⇓). Paw values were referred to atmospheric pressure, and Poes/Pga values were referred to their values at respiratory system relaxation volume (Vr); the latter values were determined during the below-described measurements of the change (Δ) in functional residual capacity (FRC). For each set of test breaths, PL was determined as the difference between the average Paw value and the averaged Poes. Total RS, chest wall, and lung inspiratory mechanical properties were computed by standard formulas (see Appendix 1).
Haemodynamics and gas exchange
In each posture, intravascular pressure transducers (Abbott, Sligo, Ireland) were set to zero at right-atrial level. Within 75–90 min after posture assumption, end-expiratory central venous and pulmonary artery wedge pressures were determined consecutively three times during respective 10-s duration ETT disconnections from the breathing circuit. ETT disconnections were separated by two 6-min duration baseline-ventilation intervals, over which heart rate, mean arterial/pulmonary artery pressure, cardiac output, and mixed-venous oxygen saturation values were recorded and averaged. Just prior to each ETT disconnection, mixed-venous and arterial blood gas samples were taken and analysed immediately (ABL System 625 blood-gas analyser model; Radiometer, Copenhagen, Denmark). After collecting the gases expired within 2 min prior to the second and third ETT disconnection, physiological dead space was determined as previously described 30. Formula-derived variables included cardiac, systemic and pulmonary vascular resistance index, oxygen consumption/delivery, oxygen extraction ratio (O2ER), right/left ventricular stroke work index (SWI), respiratory quotient, alveolar oxygen partial pressure (PO2), and shunt fraction (see Appendix 2).
ΔFRC and end-expiratory lung volume
Within 105–120 min after study-posture assumption, baseline-ventilation ΔFRC was determined twice as previously reported 32, by allowing exhalation to Vr. Expiratory V′ always reached 0 L·s−1 within <10 s. The two ΔFRC measurements were separated by 17 min of baseline ventilation. Immediately after each ΔFRC measurement, FRC was determined with the closed-circuit helium dilution technique as previously explained 30, 33–35. Helium concentration was measured with a helium-analyser (PK Morgan Ltd, Kent, UK). Helium-dilution technique limitations have been previously analysed 36. Baseline ventilation EELV was computed as the sum of measured FRC and ΔFRC.
For each posture, only the means of obtained sets of measurements were analysed. Variable comparisons were conducted with two-factor univariate ANOVA for repeated measures, followed by the Scheffé test as appropriate. Significance was accepted at p<0.05. Values are presented as mean±sd.
Full data were obtained from all study participants and no protocol-related complications occurred 37. Of the 10 patients, six were weaned from mechanical ventilation 15.2±3.1 days after its institution and discharged from the intensive care unit (ICU) after another 2–3 days and four patients died of multiple system organ failure (table 1⇓).
ΔFRC, PEEPtot of the RS and chest wall, and Pga were unaffected by body posture (table 2⇓). Poes at Vr did not differ significantly among study postures (table 2⇓). The mean maximal amplitude and mean rise rate of Poes-cardiac oscillations (fig. 1⇓), were similar (table 2⇓). Consequently, the initial, correct oesophageal balloon positioning relative to the heart was maintained throughout the study period, and Poes measurements were as accurate as possible in all study postures 23. The oesophageal balloon technique has been adopted in previous studies and is considered adequate in the prone position 30, 33, 35, 38.
Tracheal pressure, oesophageal pressure and transpulmonary pressure
Paw was unaffected by body posture (fig. 2a⇓). P1/P2 of averaged Poes tracings were higher in prone versus SRBAS/SRPP at VTs≥0.6 L (p<0.05–0.01; fig. 2b⇓). Maximal PL (Pmax,L) was lower in prone/SRPP versus SRBAS at VTs≥0.6–0.8 L (p<0.05–0.01). The P1 value of PL (P1,L) was lower in prone/SRPP versus SRBAS at VTs≥0.4–0.6 L (p<0.05–0.01). P2,L was lower in prone/SRPP versus SRBAS at VTs≥0.6 L (p<0.05–0.01; fig. 2c⇓). Pmax,L/P1,L (individual patient data not shown), and P2,L exhibited similar response patterns to posture change in all patients. Figure 3⇓ displays individual P2,LVT relationships.
EELV and VT/EELV
In prone and SRPP, baseline-ventilation EELV was significantly higher versus SRBAS (by 30.3±1.6% and 23.9±1.0%, respectively, p<0.01; table 3⇓). Accordingly, the baseline-ventilation VT (0.6 L)/EELV ratio, was lower in prone and SRPP versus SRBAS (0.41±0.04 and 0.42±0.04, respectively, versus 0.58±0.05, p<0.01). The sigh (1.0 L) VT/EELV ratio was also lower in prone and SRPP versus SRBAS (0.68±0.06 and 0.71±0.06, respectively, versus 0.97±0.09, p<0.01). VT/EELV exhibited similar response patterns to body posture in all patients.
Figure 4⇓ displays the main results for the respiratory mechanics. Static chest wall elastance was higher in prone versus SRBAS/SRPP at VTs≥0.6 L (p<0.05–0.01). Additional lung resistance was lower in prone/SRPP versus SRBAS at all employed VTs (p<0.01). Static lung elastance (Estat,L) was lower in prone/SRPP versus SRBAS at VTs≥0.6 L (p<0.05–0.01).
Haemodynamics and gas exchange
Table 4⇓ displays the main results. Directly determined haemodynamic variables were not significantly affected by posture change. The O2ER exhibited sequential significant decreases from SRBAS to SRPP. Left vetricular SWI became higher in SRPP versus SRBAS. All gas exchange variables (including Pa,O2/FI,O2, carbon dioxide arterial tension, shunt fraction, and physiological dead space) were significantly improved in prone/SRPP versus SRBAS.
The present study has shown that in severe ARDS 8, 13, 14, 20, prone positioning under PEEP optimisation reduces ventilation-induced stress (reflected by PL) and strain (reflected by VT/EELV) relative to 60° SRBAS; these effects are maintained within 2 h following return to 60° SRPP. Although only 10 patients were studied, individual response patterns of PL and VT/EELV to posture change were similar. Favourable results occur with a baseline (0.6 L) to 1.0 L VTs (range 8.2–17.6 mL·kg−1 PBW for participants in the present study). The lowest of the aforementioned VTs are similar to the 8–9 mL·kg−1VTs routinely used by physicians studying/treating patients with acute lung injury/ARDS 9, 39–42. Other pronation benefits, also maintained in SRPP, included decreased Estat,L and additional lung resistance, improved arterial oxygenation, and reduced shunt fraction and Pa,CO2 and physiological dead space; these findings are consistent with previously published results 30, 37, 43–45.
In the present study, PEEP was optimised to 4.0±0.9 cmH2O lower values relative to the pre-study PEEP. PEEP optimisation was aimed at: 1) maintaining pre-study arterial oxygenation in SRBAS; and 2) allowing for a VT increase of 1.7±0.2 mL·kg−1 PBW during the study period, without concomitant end-inspiratory stress increase to potentially injurious levels exceeding 30 cmH2O (P2,L in fig. 2c⇓) 5, 46. Posture change from near-supine to steep SRBAS may have facilitated the achievement of the aforementioned ventilation goals by partially relieving abdominal/cardiac compression of dependent/caudal lung regions 14.
Substantial lung mechanics/gas exchange benefits were obtained only after the pronation manoeuvre. The selection of the 0.6 L baseline VT probably facilitated the intratidal alveolar recruitment 30, 47; the optimised PEEP probably facilitated the maintenance of such recruitment 47. Also, the employed 1.0 L sighs were probably more effective in the prone position 48. The combined EELV and Pa,O2/FI,O2 increases and the shunt fraction decrease indicate effective re-aeration of well perfused, but previously collapsed lung units 18, 23, 30, 49. The P2,L and Estat,L reductions indicate pronation-induced reversal of regional atelectasis 23, 30. The Pa,CO2 and physiological dead space reductions suggest an increase in effective alveolar ventilation 23. The additional lung resistance decrease suggests a reduced time constant inequality range, more homogenous lung inflation, and reduced numbers of atelectatic and hyperinflated/overdistended alveoli 14, 23, 30. The reduction in regional atelectasis attenuates the stress in neighbouring lung regions 46. The reduction in regional hyperinflation or overdistension attenuates regional strain and reduces the regional probability of traumatic alveolar rupture 23. Thus, although the current authors could not directly determine regional stress/strain, the combined results in the present study strongly suggest reduced “regional probability” of VILI in the prone position.
The observed significant decreases in O2ER were due to the simultaneous decreases in oxygen consumption and/or increases in oxygen delivery. These changes could be attributable to the effects of the probably slowly increasing plasma propofol and fenanyl concentrations throughout the study period 50. Propofol/fentanyl anaesthesia reduces oxygen consumption in hypoxaemic respiratory failure patients 24. Mild and moderate maintenance infusion-induced increases in propofol and fentanyl plasma concentrations, respectively 50, should not directly affect myocardial performance 51, 52. The observed arithmetical decreases in heart rate could be attributable to fentanyl 53. The heart rate changes could have contributed to the observed arithmetical increases in SRPP cardiac index (table 4⇓) by improving cardiac diastolic filling, cardiac muscle fibre length-tension relationship, and consequently, cardiac contractility 54. The speculation for improved diastolic filling is supported by the observed increase in left ventricular SWI 55. In SRPP, the improved haemodynamic performance could have contributed to the maintenance of the pronation gas exchange benefits relative to SRBAS. However, as in the case of the prone position, the major portion of the SRPP gas exchange improvement could still be explained by the still improved shunt fraction and physiological dead space (table 4⇓), and thus, ventilation-perfusion matching.
The determinations of Poes in different body postures may be problematic, despite prior relevant methodological interpretations 30, 33, 35, 38, and confirmation of an unchanged pattern of transmission of intracardiac pressure changes to the oesophageal balloon 23. Pronation causes redistribution of regional lung collapse 56; if this occurs in lung areas adjacent to the lower oesophagus, the patterns of intrapleural pressure transmission to the oesophageal balloon may differ between prone and semirecumbent postures, thus resulting in possible posture-related bias in PL calculations. However, regional lung collapse, although different in amount, should tend to exhibit similar distribution in SRBAS and SRPP. This should have minimised the impact of the aforementioned potential problem upon the PL results in the current study, because similar improvement in gas exchange, indicating similar amount of alveolar recruitment, was accompanied by similar reductions in PL determined in prone and SRPP relative to SRBAS. Consequently, PL was probably calculated in prone as accurately as in semirecumbent posture.
Pronation benefits were maintained during SRPP measurements; major contributory factors may include: 1) optimised PEEP-induced maintenance of overall pronation-induced alveolar recruitment, despite its posture-associated redistribution 56; and 2) partial maintenance of a possible pronation-induced lower lung lobe decompression in steep semirecumbent posture 14. The maintenance of lower lobe decompression could have been facilitated by the absence of abnormally raised intra-abdominal pressure and decreased abdominal compliance (table 2⇓) 27.
According to the lung stress/strain results presented in the current study, ventilation with standard VTs (0.6 L = 9.0±0.9 mL·kg−1 PBW) 9, 39–42 in prone/SRPP is equally protective as ventilation with low VTs (0.4 L = 6.0±0.6 mL·kg−1 PBW) 25 in SRBAS. Indeed, in prone and SRPP, P2,L and Pmax,L values at baseline (0.6 L) VT were similar to SRBASP2,L and Pmax,L values at 0.4 L VT (fig. 2c⇓). VT/EELV (strain) was also similar in prone and SRPP at 0.6 L VT and SRBAS at 0.4 L VT (0.39±0.04). Consequently, if in severe ARDS, low VTs are accepted as the “gold standard” in SRBAS 12, 25, 57, pronation may provide considerable “VT-liberation capability” by allowing a 50% (3 mL·kg−1 PBW) VT increase, without appreciable increase in the VILI risk. Under PEEP optimisation, such VT liberation may be allowable in SRPP for at least 2 h.
Sighs frequently improve oxygenation in ARDS 34, 48, 58. The present authors administered 1.0-L sigh-test VTs (15.0±1.5 mL·kg−1 PBW), which resulted in significantly lower PL (fig. 3⇓) and VT/EELV in prone/SRPP. These results indicate a decreased probability of traumatic alveolar rupture, especially during frequent sigh administration (1–3·min−1) 34, 48.
Prone positioning of patients with early and severe acute respiratory distress syndrome reduces lung parenchyma stress and strain during mechanical ventilation. This suggests a reduced risk of ventilator-induced lung injury. Pronation also improves sigh safety and gas exchange efficiency. Under positive end-expiratory pressure optimisation, pronation benefits may be maintained in a post-prone semirecumbent position.
APPENDIX 1: INSPIRATORY MECHANICAL VARIABLES
For the respiratory system, chest wall, and lung the following inspiratory mechanics-variables were determined: 1) maximal, interrupter, and additional resistances, computed respectively as Pmax–P2, Pmax–P1, and P1–P2 differences divided by the preceding inspiratory flow; and 2) dynamic, and static elastances, computed as respective P1–PEEPtot and P2–PEEPtot differences divided by the administered VT. Lung interrupter resistance reflects “ohmic” airway resistance; lung additional resistance reflects lung tissue stress relaxation tension and time constant inequality.
APPENDIX 2: FORMULAS USED TO DERIVE HAEMODYNAMIC AND GAS EXCHANGE VARIABLES
Cardiac index = CO/BSA
Systemic vascular resistance index = (MAP–CVP)×80×CI−1
Pulmonary vascular resistance index = (MPAP–PAWP)×80×CI−1
Oxygen consumption per m2 BSA = CI×1.36×Hgb×(SaO2–SvO2)
Oxygen delivery per m2 BSA = CI×1.36×Hgb×Sa,O2
Oxygen extraction ratio = (Sa,O2–Sv,O2)×Sa,O2−1
SVI = CI×(heart rate)−161
Right ventricular stroke work index = (MPAP–CVP)×SVI×0.0136 61
Left ventricular stroke work index = (MAP–PAWP)×SVI×0.0136 61
Respiratory Quotient = (FEY of carbohydrate intake)×1.0+(FEY of protein intake)×0.8+(FEY of lipid intake)×0.7 62
Alveolar PO2 = Pi,O2–PA,CO2×[FI,O2–(1–FI,O2)×R−1]; Pi,O2 = FI,O2×(PB–47); PA,CO2∼Pa,CO2
O2 content of blood = Hgb×1.36×SO2×10–1+0.003×PO2
Shunt fraction = (Cc,O2–Ca,O2)×(Cc,O2–Cv,O2)–1
CO: cardiac output (L·min−1); BSA: body surface area (m2); MAP: mean arterial pressure (mmHg); CVP: central venous pressure (mmHg); 80: transformation factor of Wood units (mmHg·L−1·min) to standard metric units (dynes·s·cm−5); CI: cardiac index (L·min−1·m−2); MPAP: mean pulmonary artery pressure (mmHg); PAWP: pulmonary artery wedge pressure (mmHg); 1.36: O2 combining power of 1 g of haemoglobin (mL); Hgb: haemoglobin concentration in g·L−1; Sa,O2: arterial O2 saturation; Sv,O2: mixed venous O2 saturation; SVI: stroke volume index (mL per heart beat); 0.0136: conversion factor pressure and volume units to work units (g·m); FEY: fractional energy yield relative to total of prescribed nutritional support; Pi,O2: inspired O2 partial pressure (mmHg); PA,CO2: alveolar CO2 partial pressure (mmHg); R: respiratory quotient; PB: barometric pressure (mmHg); 47: water saturated vapour pressure at 37°C (mmHg); 0.003: O2 solubility coefficient at 37°C (mL·dL−1·mmHg); PO2: O2 partial pressure (mmHg); SO2: oxygen saturation; Cc,O2/Ca,O2/Cv,O2: O2 content in end-capillary/arterial/mixed-venous blood, respectively.
1 mmHg = 0.133 kPa.
The authors would like to thank M. Tzoufi for her valuable help in manuscript preparation.
- Received September 9, 2004.
- Accepted November 19, 2004.
- © ERS Journals Ltd