The ability to accurately assess and measure lung

A New Oxygenation Index for Reflecting Intrapulmonary Shunting in Patients Undergoing Open-Heart Surgery* Mohamad F. El-Khatib, PhD; and Ghassan W. Jamaleddine, MD Study objectives: To assess the reliability of new and traditional oxygenation measurements in reflecting intrapulmonary shunt. Design: Prospective study. Setting: Cardiac surgery unit at a university hospital. Patients: Fifty-five patients undergoing coronary artery bypass grafting. Measurements and results: Simultaneous blood samples were collected from an indwelling arterial line and a catheter for determination of blood gases. Standard accepted formulas were utilized to measure a new oxygenation index: PaO 2 /fraction of inspired oxygen (FIO 2 ) mean airway pressure (Paw). The standard formulas used were the oxygenation ratio (PaO 2 /FIO 2 ), PaO 2 /alveolar partial oxygen pressure (PAO 2 ), alveolar-arterial oxygen tension gradient (P[A-a]O 2 ), and intrapulmonary shunt (venous admixture [Qsp/Qt]). There were significant negative (p < 0.05) correlations between the PaO 2 /(FIO 2 Paw) and Qsp/Qt (r 0.85), between the PaO 2 /FIO 2 and Qsp/Qt (r 0.74), and between the PaO 2 /PAO 2 and Qsp/Qt (r 0.71). There was a significant positive (p < 0.05) correlation between the P(A-a)O 2 gradient and Qsp/Qt (r 0.66). However, the correlation was strongest between the PaO 2 /(FIO 2 Paw) and Qsp/Qt. Conclusion: In this group of patients, PaO 2 /(FIO 2 Paw) might be more reliable than other oxygenation measurements in reflecting intrapulmonary shunt. (CHEST 2004; 125:592 596) Key words: intrapulmonary shunt; mean airway pressure; open-heart surgery; oxygenation factor; oxygenation measurements; oxygenation ratio; positive end-expiratory pressure Abbreviations: ALI acute lung injury; CABG coronary artery bypass graft; Cco 2 pulmonary end-capillary oxygen content; Fio 2 fraction of inspired oxygen; P(A-a)O 2 alveolar-arterial oxygen tension gradient; Pao 2 alveolar partial oxygen pressure; Paw mean airway pressure; PEEP positive end-expiratory pressure; Qsp/Qt venous admixture The ability to accurately assess and measure lung function is essential in the management of patients requiring mechanical ventilation. Such assessments and measurements aid in diagnosis, in optimizing mechanical ventilatory support, and in predicting the likelihood of success of weaning. The Pao 2 /fraction of inspired oxygen (Fio 2 ) ratio, the Pao 2 /alveolar partial oxygen pressure (PAo 2 ) ratio, and the alveolar-arterial oxygen tension gradient (P[A-a]O 2 ) are the most common of these measurements. However, Pao 2 /Fio 2 remains the most convenient and widely used bedside index of oxygen *From the Departments of Anesthesiology (Dr. El-Khatib) and Medicine (Dr. Jamaleddine), School of Medicine, American University of Beirut, Beirut, Lebanon. Manuscript received February 10, 2003; revision accepted July 10, 2003. Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (e-mail: permissions@chestnet.org). Correspondence to: Mohamad Khatib, PhD, Department of Anesthesiology, American University of Beirut, PO Box 11-0236, Riad El Solh, Beirut 1107 2020 Lebanon; e-mail: mk05@aub.edu.lb exchange. 1 3 It was first described by Horovitz et al 4 in 1974 as an index used to compare arterial oxygenation at different levels of Fio 2. Since then, it has been commonly used to assess respiratory status as well as response to different therapies, whether the therapy is an increase in Fio 2 or changes in mechanical ventilation settings. 5 7 Moreover, Pao 2 /Fio 2 has been considered as the differentiating factor between establishing a diagnosis for acute lung injury (ALI) or a diagnosis for ARDS. 8 However, although simple to obtain, Pao 2 /Fio 2 is affected by changes in mixed venous oxygen saturation and does not remain equally sensitive across the entire range of Fio 2, especially when shunt is the major cause of admixture; another oxygenation index, Pao 2 /PAo 2, has been reported to be superior to Pao 2 /Fio 2 in this regard. 9 Also, more importantly, this ratio does not account for changes in the functional status of the lung that result from alterations in positive end-expiratory pressure (PEEP), auto- PEEP, or other techniques for adjusting average 592 Clinical Investigations in Critical Care

lung volume (ie, inverse ratio ventilation or prone positioning) during mechanical ventilation. As such, in patients receiving mechanical ventilation, Pao 2 / Fio 2 might not be a sensitive indicator particularly when assessing the severity of the lung disease or when tracking the oxygen-exchanging status of the lung is desired in the presence of such interventions as PEEP and prone positioning. The Pao 2 /(Fio 2 Paw ratio, a new oxygenation index termed oxygenation factor, which is based on the usual Pao 2 /Fio 2 but takes into consideration some important mechanical ventilatory support variables such as PEEP, inspiratory time fraction, and tidal volume, could be a better and superior indicator for assessing the severity of disease and/or for tracking the oxygen-exchanging status. The aim of this study is to assess the reliability of this oxygenation factor and other oxygenation measurements in reflecting intrapulmonary shunt in patients following open-heart surgery. Materials and Methods This study was approved by the Institutional Review Board, and a consent was obtained prior to the initiation of the study. Fifty-five hemodynamically and clinically stable patients receiving mechanical ventilation in the cardiac surgery unit following coronary artery bypass graft (CABG) surgery were included in the study. These were consecutive patients in whom CABG surgery was performed with a cardiopulmonary bypass pump. All patients were monitored with continuous electrocardiography, BP, and pulse oximetry during the whole study. All patients were receiving mechanical ventilation (PB-7200ae; Puritan-Bennett, Mallinckrodt; St. Louis, MO). As per routine monitoring, all patients had a Swan-Ganz catheter and an indwelling arterial line. A period of at least 10 min was allowed before data collection, during which the patients were not disturbed by nursing procedures and were not disconnected from the ventilator for suctioning. Within the first hour of admission to the cardiac surgery unit, simultaneous blood samples from the arterial line and the distal and proximal ports of the Swan-Ganz catheter (Arrow; Reading, PA) were obtained and immediately subjected to duplicate blood gas analysis in two separate blood gas machines (ABL-720 and ABL-520; Radiometer; Copenhagen, Denmark). All blood samples were collected using the same model and brand of syringes (Preset Vacutainer System; Becton-Dickinson; Plymouth, UK). Blood samples were not obtained more than twice for any one patient. From blood gas measurements, Pao 2 /(Fio 2 Paw), Pao 2 /Fio 2,Pao 2 /PAo 2, P(A-a)O 2, and the intrapulmonary shunt (venous admixture [Qsp/Qt]) were determined. Qsp/Qt was determined from the following formula: (Cco 2 arterial oxygen content)/(cco 2 mixed venous oxygen content), where Cco 2 end-pulmonary capillary oxygen content. 10 Mean values and SD were calculated for each variable. Standard techniques of linear regression and correlation by the least-square method were used to assess the degree of correlation among these variables. Student t test was used for statistical analysis of the correlation coefficients. Statistical significance was considered at the 5% level (p 0.05). Variables Table 1 Demographics* Results Patients characteristics are presented in Table 1. A total of 74 sets of data were obtained from 55 patients due to the fact that 1 extra set of data were obtained from 19 patients following changes in their PEEP and/or Fio 2, which were clinically indicated and under the discretion of the medical team who were blinded to the study except for the arterial blood gas values. There was significant negative linear relationships between Qsp/Qt and Pao 2 /(Fio 2 Paw (r 0.85, p 0.05), between Qsp/Qt and Pao 2 /Fio 2 (r 0.74, p 0.05), and between Qsp/Qt and Pao 2 /PAo 2 (r 0.71, p 0.05) [Figs 1 3]. However, there was a significant positive linear relationship between Qsp/Qt and P(A-a)O 2 gradient (r 0.66, p 0.05) [Fig 4]. As shown in Table 2, the correlation was strongest between Qsp/Qt and Pao 2 /Fio 2 Paw). Discussion Data Patients 55 Age, yr 61 6 Male/female gender 36/19 Weight, kg 74 5 Height, cm 167 4 Smokers/ex-smokers/nonsmokers 7/28/20 *Data are presented as No. or mean SD. Our data demonstrate that the new oxygenation index (Pao 2 /Fio 2 Paw), the oxygenation ratio Figure 1. Correlation between intrapulmonary shunt (Qsp/Qt) and new oxygenation index (Pao 2 /Fio 2 Paw). Regression line: slope 0.38, intercept 18.83. www.chestjournal.org CHEST / 125 / 2/ FEBRUARY, 2004 593

Figure 2. Correlation between intrapulmonary shunt (Qsp/Qt) and Pao 2 /Fio 2 ratio. Regression line: slope 0.05, intercept 21.40. Figure 4. Correlation between the intrapulmonary shunt (Qsp/ Qt) and P(A-a)O 2. Regression line: slope 0.04, intercept 0.4. (Pao 2 /Fio 2 ), Pao 2 /PAo 2, and P(A-a)O 2 are reliable reflectors of intrapulmonary shunt (Qsp/Qt). However, in this group of patients, Pao 2 /Fio 2 Paw is superior to other oxygenation measurements in reflecting intrapulmonary shunt. The intrapulmonary shunt fraction index has been considered the gold standard for the clinical assessment of lung oxygenation function. 11 15 This index most accurately represent the lung oxygenation function when direct measurements of both arterial and pulmonary arterial blood samples are available as well as when the Fio 2 is consistent. 11 Indexes of arterial oxygenation, such as Pao 2 /Fio 2, Pao 2 /PAo 2, P(A-a)O 2, and Pao 2 /Fio 2 Paw, which do not require mixed venous blood samples, are useful since these indexes can be applied to patients regardless of whether a pulmonary artery catheter is in place. 16 However, Pao 2 /Fio 2 remains the mostly used and evaluated index due to its simplicity and the fact that it can be determined in both spontaneously breathing patients and patients receiving mechanical ventilation. There are conflicting data on the accuracy with which Pao 2 /Fio 2 reflects oxygen exchange. Pao 2 / Fio 2 values 200 mm Hg have been reported to correlate well with Qsp/Qt of 20%. 17,18 Gowda and Klocke 13 have shown that Pao 2 /Fio 2 is a useful estimation of the degree of gas exchange abnormality under usual clinical conditions. Also two studies 17,19 have demonstrated that Pao 2 /Fio 2 correlated closely with measured venous admixture, especially in hemodynamically stable patients, and that this ratio correlated better than any other tension-based oxy- Table 2 Comparison of Oxygenation Indices to Intrapulmonary Shunt* Variables Mean SD r Value Figure 3. Correlation between intrapulmonary shunt (Qsp/Qt) and Pao 2 /Pao 2 ratio. Regression line: slope 26.97, intercept 19.75. Qsp/Qt 8.8 5.1 1.00 Pao 2 27 11 0.85 Pao 2 /Fio 2 231 69 0.74 Pao 2 /Pao 2 0.41 0.13 0.71 P(A-a)O 2 198 78 0.66 Fio 2 48 11 N/A Paw 10 3 N/A *N/A not applicable. 594 Clinical Investigations in Critical Care

genation index. However, more recent data 15 suggested Pao 2 /Fio 2 is no substitute for Qsp/Qt during mechanical ventilation, and suggested that the use of Pao 2 /Fio 2 can result in misclassification on the gas exchange scale suggested by the American-European Consensus Conference on ARDS. 8 Although Pao 2 /PAo 2 and P(A-a)O 2 have the advantage for use in both spontaneously breathing patients and patients receiving mechanical ventilation, applying the alveolar gas equation to calculate PAo 2 involves several assumptions: (1) that PAco 2 equals Paco 2, (2) that the respiratory exchange ratio equals 0.8, and (3) that partial pressure of water in alveolar gas equals 47 mm Hg. Furthermore, the knowledge of barometric pressure is also essential in determining PAo 2. The assumptions listed above could explain why the appropriateness of P(A-a)O 2 and Pao 2 /PAo 2 in quantifying disruption in pulmonary oxygen transfer have been questionable in previous studies. 9,11,20,21 Pao 2 /Fio 2 has also been widely used for patients receiving mechanical ventilation. It has been used for both tracking responses to changes in ventilatory support 1,6,7 as well as in classification of the severity of lung disease. 7,8 However, Pao 2 /Fio 2 does not account for the functional status of the lung, primarily for changes in end-expiratory lung volume due to changes in PEEP and/or auto-peep. As such, when used to track responses to changes in ventilatory support or to classify patients into the ALI group or the ARDS group, the Pao 2 /Fio 2 ratio per se might be misleading. For instance, two mechanically ventilated patients with identical Pao 2 /Fio 2 values but with different PEEP values should not be classified or evaluated similarly. As such, reporting Pao 2 /Fio 2 without reflecting the level of PEEP might be misleading. In the current study, we have elected to add the Paw rather than PEEP to Pao 2 /Fio 2 for several reasons. First, Paw does incorporate the effect of PEEP. Second, Paw also incorporates the effect of inspiratory and expiratory times. Finally, Paw incorporates the effect of tidal volume and/or peak inspiratory pressure depending on the mode of mechanical ventilation. These variables are all important in contributing to the lung volumes and thus the lung oxygenation function. Because it incorporates the effect of Paw into Pao 2 /Fio 2, Pao 2 /Fio 2 Paw might be a superior reflection of the gas exchange status and lung function when patients are receiving mechanical ventilation. Changes in functional status of the lung that result from alterations in PEEP, auto-peep, or other techniques for adjusting average lung volume (ie, inverse-ratio ventilation or prone positioning) are better reflected in the Pao 2 /Fio 2 Paw ratio via their effects on Paw. 10 In its new form, Pao 2 /Fio 2 Paw remains easy to determine and to apply at the bedside. In contrast to the shunt fraction, which requires the use of a Swan-Ganz catheter, the new oxygenation index is noninvasive and can be determined with a simple arterial puncture. For the determination of Pao 2 /Fio 2 Paw, the value for the Paw, which is readily available on most modern mechanical ventilators, is needed in addition to the Pao 2 and Fio 2 values. Through the inclusion of the Paw, Pao 2 /Fio 2 Paw) will account for changes in the functional status of the lung resulting from use and alterations in PEEP and should be useful in the clinical practice. Despite these advantages, one limitation of the new oxygenation index remains that it cannot be used on nonintubated and spontaneously breathing patients since the index will be undefined as Paw will be zero under these conditions. In the current study, the patient population received mechanical ventilation for postoperative management following open-heart CABG surgeries. The intrapulmonary shunt in our patients ranged from 2 to 20%. This range of shunt fraction does not reflect severe gas exchange disorders. However, the use of PEEP on patients after open-heart surgeries has been shown to increase end-expiratory lung volume, 22 to result in fewer atelectatic and fewer unperfused lung units, 23 and to improve oxygenation. 24 27 Pao 2 /Fio 2 Paw, which incorporates PEEP (through the Paw), should better reflect lung function than other oxygenation indexes that do not incorporate PEEP. Furthermore, the aim of the current study was only to assess the reliability of Pao 2 /Fio 2 Paw in reflecting intrapulmonary shunt and not to identify a cut-off or a threshold value that will differentiate between clinically acceptable and unacceptable intrapulmonary shunt fractions. For this purpose, the current study was conducted in a group of postoperative CABG patients in whom arterial lines as well as a Swan-Ganz catheters are placed for routine monitoring, which will allow determination of the different oxygenation indexes. We still need to validate the current findings in patients receiving mechanical ventilation for ALI or ARDS for whom a wider and higher range of intrapulmonary shunt fraction would be expected. Conclusion In conclusion, our data show that the currently used oxygenation measurements can be used to reflect intrapulmonary shunt in patients following open-heart surgery. However, a new and simple oxygenation index, Pao 2 /Fio 2 Paw, might be suwww.chestjournal.org CHEST / 125 / 2/ FEBRUARY, 2004 595

perior to most common oxygenation indexes in this group of patients. Further studies are needed to evaluate any role of Pao 2 /Fio 2 Paw in assessing and following up lung function in patients with ALI or ARDS. ACKNOWLEDGMENT: We thank the staff of the Department of Inhalation Therapy and the nursing staff in the Cardiac Surgery Unit. References 1 The Acute Respiratory Distress Syndrome Network. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. N Engl J Med 2000; 342:1301 1308 2Räsänen J, Downs JB, Malec DJ, et al. Oxygen tensions and oxyhemoglobin saturations in the assessment of pulmonary gas exchange. Crit Care Med 1987; 15:1058 1061 3 Gould MK, Ruoss SJ, Rizk NW, et al. Indices of hypoxemia in patients with acute respiratory distress syndrome: reliability, validity, and clinical usefulness. Crit Care Med 1997; 25:6 8 4 Horovitz JH, Carrico CJ, Shires GT. Pulmonary response to major injury. Arch Surg 1974; 108:349 355 5 Baigelman W, Bellin SJ, Pearce L, et al. Relation of inspired oxygen fraction to hypoxemia in mechanically ventilated adults. Crit Care Med 1984; 12:486 488 6 Sydow M, Burchardi H, Ephraim E, et al. Long-term effects of two different ventilatory modes on oxygenation in acute lung injury: comparison of airway pressure release ventilation and volume-controlled inverse ratio ventilation. Am J Respir Crit Care Med 1994; 149:1550 1556 7 Bone RC, Maunder R, Slotman G, et al. An early test of survival in patients with the adult respiratory distress syndrome: the Pao 2 /Fio 2 ratio and its differential response to conventional therapy; Prostaglandin E 1 Study Group. Chest 1989; 96:849 851 8 Bernard GR, Artigas A, Brigham KL, et al. The American- European Consensus Conference on ARDS: definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med 1994; 149:818 824 9 Chatburn R, Lough M. Handbook of respiratory care. 2nd ed. Chicago, IL: Year Book Medical Publishers, 1990 10 Cane RD, Shapiro BA, Templin R, et al. Unreliability of oxygen tension-based indices in reflecting intrapulmonary shunting in critically ill patients. Crit Care Med 1988; 16: 1243 1245 11 Nunn JF. Applied respiratory physiology. 4th ed. Oxford UK: Butterworth-Heinemann, 1993 12 Siggaard-Anderson O, Gothgen H. Oxygen parameters of arterial and mixed venous blood: new and old. Acta Anaesthesiol Scand 1995; 39:41 46 13 Gowda MS, Klocke RA. Variability of indices of hypoxemia in adult respiratory distress syndrome. Crit Care Med 1997; 25:41 45 14 Nirmalan M, Willard T, Columb MO, et al. Effect of changes in arterial-mixed venous oxygen content difference (C(a-v)O 2 ) on indices of pulmonary oxygen transfer in a model ARDS lung. Br J Anaesth 2001; 86:477 485 15 Murray JF, Matthay MA, Luce JM, et al. An expanded definition of the adult respiratory distress syndrome. Am Rev Respir Dis 1988; 138:720 723 16 Covelli HD, Nessan VJ, Tuttle WK. Oxygen derived variables in acute respiratory failure. Crit Care Med 1983; 11:646 649 17 Robinson NB, Weaver LJ, Carrico CJ, et al. Evaluation of pulmonary dysfunction in the critically ill [abstract]. Am Rev Respir Dis 1981; 123:92 18 Zetterstorm H. Assessment of the efficiency of pulmonary oxygenation: the choice of oxygenation index. Acta Anaesthesiol Scand 1988; 32:579 584 19 Hoffstein V, Duguid N, Zamel N, et al. Estimation of changes in alveolar-arterial oxygen gradient induced by hypoxia. J Lab Clin Med 1984; 104:685 692 20 Hess D, Maxwell C. Which is the best index of oxygenation: P(A-a)O 2, Pao 2 /PAo 2, or Pao 2 /Fio 2? Respir Care 1985; 30:961 963 21 Viale JP, Percival CJ, Annat G, et al. Arterial-alveolar oxygen partial pressure ratio: a theoretical reappraisal. Crit Care Med 1986; 14:153 154 22 Valta P, Takala J, Elissa T, et al. Effects of PEEP on respiratory mechanics after open heart surgery. Chest 1992; 141:281 289 23 Downs JB, Mitchell LA. Pulmonary effects of ventilatory pattern following cardiopulmonary bypass. Crit Care Med 1976; 4:295 300 24 Marvel SL, Elliott CG, Tocino I, et al. Positive end-expiratory pressure following coronary artery bypass grafting. Chest 1986; 90:537 541 25 Suter PM, Demottaz V, Hemmer M. Ventilatory support after open heart surgery: effects of PEEP and CPAP on gas exchange and lung function. Herz 1978 3:198 205 26 Berthelsen P, St Haxholdt O, Husum B, et al. PEEP reverses nitroglycerin-induced hypoxemia following coronary artery bypass surgery. Acta Anaesthesiol Scand 1986; 30:243 246 27 Dyhr T, Laursen N, Larsson A. Effects of lung recruitment maneuver and positive end-expiratory pressure on lung volume respiratory mechanics and alveolar gas mixing in patients ventilated after cardiac surgery. Acta Anaesthesiol Scand 2002; 46:717 725 596 Clinical Investigations in Critical Care