Tracking Respiratory Therapy in the Trauma Patient Michael A. Goldfarb, MD, MAJ, MC, Edgewood Arsenal, Aberdeen Proving Ground, Maryland Terrence F. Ciurej, MD, CAPT, MC, Edgewood Arsenal, Aberdeen Proving Ground, Maryland 1. C. McAslan, MB, Baltimore, Maryland William J. Sacco, PhD, Baltimore, Maryland Michael A. Weinstein, MD, MAJ, MC, Edgewood Arsenal, Aberdeen Proving Ground, Maryland R. A. Cowley, MD, Baltimore, Maryland
The purpose of this research was to test a respiratory index (RI) as an indicator of a trauma patient’s respiratory state. If a patient’s respiratory state could be simply characterized and followed, this would allow: (1) comparison of therapy in patients with respiratory complications in various institutions, (2) comparison of variations in treatment, and (3) graphic representation of improvement or deterioration in a patient’s condition as an adjunct to patient care by means of a nomogram described subsequently herein. An increase of the alveolar-arterial oxygen difference can be an indicator of hypoxemia and is an important consideration in controlling arterial oxygenation in the clinical environment. The alveolar-arterial oxygen difference results from venous admixture (or physiologic shunt) which is caused by shunted venous blood that mixes with oxygenated blood leaving the pulmonary capillaries and by uneven ventilation-perfusion ratios in different parts of the lung [1,2].
From the Biophysics Division, Biomedical Laboratory, Edgewood Arsenal, Aberdeen Proving Ground, Maryland, and the Maryland Institute for Emergency Medicine, Saftimore. Maryland. This work was supported by Projects lE762708A090 (EA), 30873 (LWL), and DAAD0573C0032 (AMSAA). Reprint requests should be addressed to Michael A. Goldfarb. MD, Monmouth Medical Center, 225 Third Avenue, Long Branch, New Jersey 07740.
volume 129, March 1975
The following equation was suggested to the investigators by Siegel [3] and Siegel and Farrell [4]. The numerator reflects the alveolar-arterial oxygen difference. [(PH - P,, oT)F,,,_ - PaCO,] - PaO, P(AaD0,) = PaQ PaO, where Pu = barometric pressure PuzoT = alveolar water vapor pressure at the patient’s temperature (T) (approximately 47 mm I-k)
PaCOs = arterial partial pressure of carbon dioxide assumed to be equal to the alveolar partial pressure of the carbon dioxide (PaCO2) (51 Fro? = fractional concentration of oxygen in inspired gas Pa02 = arterial partial pressure of oxygen. It is possible to calculate the same P(AaDOs) for two clinically different situations. The P(AaDO2) was therefore divided by the Pa02 to derive a quotient that more accurately reflects the clinical state. For example, one patient may be on 70 per cent 0s and have a PaOz of 70 mm Hg to give a P(AaD02) of about 400 mm Hg. Another patient on 90 per cent 02 may have a Pa02 of 200 mm Hg to give a P(AaD02) also of 400 mm Hg. If the
255
Goldfarb et al
TABLE I
Frequency Index
Maximal Respiratory lndex(RI)
for
Maximal
Survivals
o-1 1.1-2 2.1-3 3.1-4 4.1-5 5.1-6 6.1 Total
TABLE II
Charts
Respiratory
Deaths
Number
Percent
Number
Percent
33 23 21 16 12 8 3 116
28.4 19.8 18.1 13.8 10.3 6.9 2.1
2 6 8 6 7 7 25 61
3.3 9.8 13.1 9.8 11.5 11.5 40.9
100%
100%
Maximal Respiratory Indexes and Probabilities of Survival
Maximal Respiratory lndex(RI) o-1 1.1-2 2.1-3 3.1-4 4.1-5 5.1-6 >5 >6
Probability of Survival 0.95 0.8 0.73 0.73 0.64 0.55 0.26 0.12
P(AaD02) in each case is divided by the PaOn, the RI is about 6 in the former and 2 in the latter. The RI thus differentiates the severity between the two conditions whereas the P(AaD02) alone does not. Since the P(AaD02) reflects shunting, the RI would also reflect pulmonary shunting. This report will indicate the usefulness of the RI when a patient’s respiratory pathophysiologic condition involves shunting. Some of the more common clinical conditions include pulmonary contusion, pulmonary embolus, and atelectasis caused by bronchial obstruction, mechanical compression of the lung, or hypoventilation [5]. The most common catise in this series was the respiratory distress syndrome of trauma. Method The patients evaluated were all treated at the Maryland Institute for Emergency Medicine. A computerized data bank allowed the selection of a group of consecutive, intubated patients in the unit at least one day. The individual charts were then reviewed and the maximal RI for the patient’s entire course was recorded.
Most patients were victims of trauma and were delivered by helicopter from the scene of the accident which occurred between March 1971 and August 1972. About 32 per cent represent transfers from other hospitals at
256
begun. Of the primary types of insults, 71.8 per cent incurred blunt trauma (automobile accidents), 7.3 per cent gunshot wounds, 2.8 per cent burns, 5.1 per cent elective surgical problems, and 13.0 per cent medical problems. Of the group with trauma, the major areas affected included the central nervous system in 27.4 per cent, thorax in 25.4 per cent, abdomen in 17.9 per cent, musculoskeletal system in 17.5 per cent, and head in 11.9 per cent. Many trauma patients had multiple injuries. Each patient was intubated at least once during his hospital course and placed on the Engstrom@ respirator. The Engstrom respirator was used in every patient in whom mechanical ventilation was required. Since the patients were on Engstrom respirators, the per cent of oxygen administered could be calculated by plotting the liters per minute of oxygen against the total minute volume in liters per minute. This nomogram accompanies all Engstrom respirators. The Engstrom respirator was calibrated each time before its use and varied about f5 per cent in accuracy. Arterial blood gases were measured every time in therapy or when the F102 was altered. Arterial blood gases were also measured at least every six hours in every intubated patient. leastone day aftertreatmentwas
Results and Comments Respiratory Index Data and Associated Probabilities of Survival. The accompanying retrospective statistics in Table I represeht a total of 177 patients, 116 of whom lived and 61 died. According to Mellemgaard, acceptable upper limits of alveolar-arterial differences (on room air) for ages twenty, forty, and sixty are 19, 24, and 28 mm Hg, respectively. Associated values of Pa02 (on room air) for the same age groups are 85,80, and 75 mm Hg. If the alveolar-arterial difference is divided by the PaOz, the respiratory indexes are 0.22,0.3, and 0.37, respectively [6]. In every case, when the RI was as high as 2, the patient was noted to be intubated. This is understandable since an RI of 2, with the patient on room air (20 per cent oxygen), and a PaC02 of 35 mm Hg would mean that the patient would have a Pa02 of 37 mm Hg. The RIs listed represent the maximal RIs of the patients’ entire hospital stay. (Table I.) The RI probabilities are generated from the group of patients treated at the Maryland Institute for Emergency Medicine. Each group of patients with maximal RIs has an associated probability of survival. The probabilities should be expected to change somewhat depending on the therapy at a particular institution. There was only one patient who lived with an RI over 7. This patient had an RI of 10.5 for a brief time subsequent to a bilateral pneumothorax sec-
The American Journal of Surgery
Tracking Respiratory
Therapy
460 440 420 400 300 360 f
340
g
320
g
300
I
60
Flop -
PERCENT
70 INSPIRED
60
90
100
OXYGEN
Figure 1. Respiratory index nomogram.
ondary to positive end expiratory pressure (PEEP). Once the bilateral pneumothorax was treated, the patient’s RI slowly returned to normal. Eight patients survived with an RI between 5 and 6, and only two survived with an RI between 6 and 7. Thus, eleven patients survived with an RI over 5. However, thirty-two of the sixty-one patients who died had an RI of 5 or greater. The probability of survival with an RI over 5 was 26 per cent, and an RI over 6 was associated with a 12 per cent probability of survival. Other survival probabilities are listed in Table II. It should be stressed that the reliability of the RI is enhanced by its irreversibility. For example, even if the RI is over 5 and returns to normal with treatment over the next several days, the patients appear to have the same poor probability of survival as if the RI had remained at 5. This became evident when it was noted that in half the patients with an RI greater than 5, the highest measurements were recorded at least one day prior to the day of death.
Volume
128,
March
1975
The other half of the patients on the day of death.
had the highest RI
Respiratory Index Nomogram. To graphically track respiratory therapy for this group of 177 patients, a nomogram was designed. (Figure 1.) Respiratory index isobars were calculated, varying Pa02 and per cent of 02 administered but keeping a constant PaCOz of 35 mm Hg. The variations in RI with a PaCOz of 20 or 60 mm Hg instead of 35 mm Hg have been calculated. For example, if the patient is on 40 per cent 02, the Pa02 is 100 mm Hg, and the PaCOY ranges from 20 to 35 to 60 mm Hg, the corresponding RIs are 0.3, 0.15, and 0.10. This variation decreases as the RI increases. In general, 0.2 RI units can be added for each 15 mm Hg decrease in PaC02 from a PaCOz of 35 mm Hg and 0.2 RI units subtracted for every 15 mm Hg increase in PaC02 above 35 mm Hg. It should be noted that there is minimal increase in Pa02 once a patient is on the RI 3 isobar. Here, if the patient is on 60 per cent 02 and the Pa02 is 90 mm Hg,
257
Goldfarb et al
placing the patient on 100 per cent 0s would give a Pa02 of 160 mm Hg. The amount of increase in the Pa02 decreases as the slopes of the isobars decrease. This reflects the generally accepted notion that if a patient is sufficiently ill to require 60 per cent Fro*, increasing the Fro:! beyond 60 per cent does not effect much of an increase in PaOz. In addition, the risk of oxygen toxicity is a hazard when the patient is placed on high concentrations for long periods [2]. Conclusions
There are several areas of research that require further investigation with regard to the respiratory index. They include: 1. Testing the RI with data from trauma patients in a prospective fashion. Subgroups of trauma patients may clarify the prognostic significance of the RI with regard to cases of isolated thoracic trauma, multiple trauma including the thorax, and multiple trauma excluding the thorax. The relevance of the RI in pulmonary units (nonsurgical) should prove helpful in tracking therapy. 2. The applicability of the nomogram will be tested further by the physicians handling respiratory problems as a graphic representation of the influence of various forms of therapy on respiratory pathophysiology. Various therapeutic regimens can be compared with similar groups of patients. 3. The prognostic and therapeutic value of the RI must be compared with more complex pulmonary indexes such as per cent of pulmonary shunting, pulmonary compliance, and functional residual capacity. Such indexes may perhaps be added to the RI so that respiratory pathophysiology can be characterized and collated even more precisely. The RI alone, however, provides a simple way to record data in a form that is easily understood.
258
Summary
The respiratory index (RI), P(AaDOz)/PaOz, was investigated in a retrospective study of 1’77 intubated patients treated at the Maryland Institute for Emergency Medicine. An RI of 0.1 to 0.37 is normal. Patients with an RI of 2 or greater were intubated. Those patients who reached an RI of 6 or more had an associated 12 per cent probability of survival. The RI reflects the presence of pulmonary shunting in a variety of circumstances including atelectasis, pulmonary contusion, and pulmonary emboli. A nomogram that allows the course of the patient with respiratory problems to be followed is described. Movement along the same isobars or between isobars can be followed by plotting the Pa02 against the Fro,. Thus, the rationale and effect of respiratory therapeutic manipulations may be recorded graphically. Acknowledgment: We wish to thank Mr Paul H. Broome for computerization of the data given herein and MS Marion Royston for technical assistance in the preparation of this report; both are from the Biomedical Laboratory, Edgewood Arsenal, Aberdeen Proving Ground, Maryland. References Nunn JF: Applied Respiratory Physiology. London, Butterworths. 1969, p 337. Pontoppidan H, Geffin B, Lowenstein E: Acute Respiratory Failure in the Adult. N Eng/ J Med 267: 743, 1972. Siegel JH: Personal communication. Siegel JH, Farrell EJ: A computer simulation model to study the clinical observability of ventilation and perfusion abnormalities in human shock states. Surgery 73: 898, 1973. 5. Laver MB, Austen WG: Cardiorespiratory dynamics, chapt 1, p 2. Surgery, 2nd ed. Philadelphia, Saunders, 1969. 6. Mellemgaard K: The alveolar-arterial oxygen difference: its size and components in normal man. Acta Pbysb/ Stand 73: 10, 1966.
The American Journal ol Surgery
The purpose of this research was to test a respiratory index (RI) as an indicator of a trauma patient’s respiratory state. If a patient’s respiratory state could be simply characterized and followed, this would allow: (1) comparison of therapy in patients with respiratory complications in various institutions, (2) comparison of variations in treatment, and (3) graphic representation of improvement or deterioration in a patient’s condition as an adjunct to patient care by means of a nomogram described subsequently herein. An increase of the alveolar-arterial oxygen difference can be an indicator of hypoxemia and is an important consideration in controlling arterial oxygenation in the clinical environment. The alveolar-arterial oxygen difference results from venous admixture (or physiologic shunt) which is caused by shunted venous blood that mixes with oxygenated blood leaving the pulmonary capillaries and by uneven ventilation-perfusion ratios in different parts of the lung [1,2].
From the Biophysics Division, Biomedical Laboratory, Edgewood Arsenal, Aberdeen Proving Ground, Maryland, and the Maryland Institute for Emergency Medicine, Saftimore. Maryland. This work was supported by Projects lE762708A090 (EA), 30873 (LWL), and DAAD0573C0032 (AMSAA). Reprint requests should be addressed to Michael A. Goldfarb. MD, Monmouth Medical Center, 225 Third Avenue, Long Branch, New Jersey 07740.
volume 129, March 1975
The following equation was suggested to the investigators by Siegel [3] and Siegel and Farrell [4]. The numerator reflects the alveolar-arterial oxygen difference. [(PH - P,, oT)F,,,_ - PaCO,] - PaO, P(AaD0,) = PaQ PaO, where Pu = barometric pressure PuzoT = alveolar water vapor pressure at the patient’s temperature (T) (approximately 47 mm I-k)
PaCOs = arterial partial pressure of carbon dioxide assumed to be equal to the alveolar partial pressure of the carbon dioxide (PaCO2) (51 Fro? = fractional concentration of oxygen in inspired gas Pa02 = arterial partial pressure of oxygen. It is possible to calculate the same P(AaDOs) for two clinically different situations. The P(AaDO2) was therefore divided by the Pa02 to derive a quotient that more accurately reflects the clinical state. For example, one patient may be on 70 per cent 0s and have a PaOz of 70 mm Hg to give a P(AaD02) of about 400 mm Hg. Another patient on 90 per cent 02 may have a Pa02 of 200 mm Hg to give a P(AaD02) also of 400 mm Hg. If the
255
Goldfarb et al
TABLE I
Frequency Index
Maximal Respiratory lndex(RI)
for
Maximal
Survivals
o-1 1.1-2 2.1-3 3.1-4 4.1-5 5.1-6 6.1 Total
TABLE II
Charts
Respiratory
Deaths
Number
Percent
Number
Percent
33 23 21 16 12 8 3 116
28.4 19.8 18.1 13.8 10.3 6.9 2.1
2 6 8 6 7 7 25 61
3.3 9.8 13.1 9.8 11.5 11.5 40.9
100%
100%
Maximal Respiratory Indexes and Probabilities of Survival
Maximal Respiratory lndex(RI) o-1 1.1-2 2.1-3 3.1-4 4.1-5 5.1-6 >5 >6
Probability of Survival 0.95 0.8 0.73 0.73 0.64 0.55 0.26 0.12
P(AaD02) in each case is divided by the PaOn, the RI is about 6 in the former and 2 in the latter. The RI thus differentiates the severity between the two conditions whereas the P(AaD02) alone does not. Since the P(AaD02) reflects shunting, the RI would also reflect pulmonary shunting. This report will indicate the usefulness of the RI when a patient’s respiratory pathophysiologic condition involves shunting. Some of the more common clinical conditions include pulmonary contusion, pulmonary embolus, and atelectasis caused by bronchial obstruction, mechanical compression of the lung, or hypoventilation [5]. The most common catise in this series was the respiratory distress syndrome of trauma. Method The patients evaluated were all treated at the Maryland Institute for Emergency Medicine. A computerized data bank allowed the selection of a group of consecutive, intubated patients in the unit at least one day. The individual charts were then reviewed and the maximal RI for the patient’s entire course was recorded.
Most patients were victims of trauma and were delivered by helicopter from the scene of the accident which occurred between March 1971 and August 1972. About 32 per cent represent transfers from other hospitals at
256
begun. Of the primary types of insults, 71.8 per cent incurred blunt trauma (automobile accidents), 7.3 per cent gunshot wounds, 2.8 per cent burns, 5.1 per cent elective surgical problems, and 13.0 per cent medical problems. Of the group with trauma, the major areas affected included the central nervous system in 27.4 per cent, thorax in 25.4 per cent, abdomen in 17.9 per cent, musculoskeletal system in 17.5 per cent, and head in 11.9 per cent. Many trauma patients had multiple injuries. Each patient was intubated at least once during his hospital course and placed on the Engstrom@ respirator. The Engstrom respirator was used in every patient in whom mechanical ventilation was required. Since the patients were on Engstrom respirators, the per cent of oxygen administered could be calculated by plotting the liters per minute of oxygen against the total minute volume in liters per minute. This nomogram accompanies all Engstrom respirators. The Engstrom respirator was calibrated each time before its use and varied about f5 per cent in accuracy. Arterial blood gases were measured every time in therapy or when the F102 was altered. Arterial blood gases were also measured at least every six hours in every intubated patient. leastone day aftertreatmentwas
Results and Comments Respiratory Index Data and Associated Probabilities of Survival. The accompanying retrospective statistics in Table I represeht a total of 177 patients, 116 of whom lived and 61 died. According to Mellemgaard, acceptable upper limits of alveolar-arterial differences (on room air) for ages twenty, forty, and sixty are 19, 24, and 28 mm Hg, respectively. Associated values of Pa02 (on room air) for the same age groups are 85,80, and 75 mm Hg. If the alveolar-arterial difference is divided by the PaOz, the respiratory indexes are 0.22,0.3, and 0.37, respectively [6]. In every case, when the RI was as high as 2, the patient was noted to be intubated. This is understandable since an RI of 2, with the patient on room air (20 per cent oxygen), and a PaC02 of 35 mm Hg would mean that the patient would have a Pa02 of 37 mm Hg. The RIs listed represent the maximal RIs of the patients’ entire hospital stay. (Table I.) The RI probabilities are generated from the group of patients treated at the Maryland Institute for Emergency Medicine. Each group of patients with maximal RIs has an associated probability of survival. The probabilities should be expected to change somewhat depending on the therapy at a particular institution. There was only one patient who lived with an RI over 7. This patient had an RI of 10.5 for a brief time subsequent to a bilateral pneumothorax sec-
The American Journal of Surgery
Tracking Respiratory
Therapy
460 440 420 400 300 360 f
340
g
320
g
300
I
60
Flop -
PERCENT
70 INSPIRED
60
90
100
OXYGEN
Figure 1. Respiratory index nomogram.
ondary to positive end expiratory pressure (PEEP). Once the bilateral pneumothorax was treated, the patient’s RI slowly returned to normal. Eight patients survived with an RI between 5 and 6, and only two survived with an RI between 6 and 7. Thus, eleven patients survived with an RI over 5. However, thirty-two of the sixty-one patients who died had an RI of 5 or greater. The probability of survival with an RI over 5 was 26 per cent, and an RI over 6 was associated with a 12 per cent probability of survival. Other survival probabilities are listed in Table II. It should be stressed that the reliability of the RI is enhanced by its irreversibility. For example, even if the RI is over 5 and returns to normal with treatment over the next several days, the patients appear to have the same poor probability of survival as if the RI had remained at 5. This became evident when it was noted that in half the patients with an RI greater than 5, the highest measurements were recorded at least one day prior to the day of death.
Volume
128,
March
1975
The other half of the patients on the day of death.
had the highest RI
Respiratory Index Nomogram. To graphically track respiratory therapy for this group of 177 patients, a nomogram was designed. (Figure 1.) Respiratory index isobars were calculated, varying Pa02 and per cent of 02 administered but keeping a constant PaCOz of 35 mm Hg. The variations in RI with a PaCOz of 20 or 60 mm Hg instead of 35 mm Hg have been calculated. For example, if the patient is on 40 per cent 02, the Pa02 is 100 mm Hg, and the PaCOY ranges from 20 to 35 to 60 mm Hg, the corresponding RIs are 0.3, 0.15, and 0.10. This variation decreases as the RI increases. In general, 0.2 RI units can be added for each 15 mm Hg decrease in PaC02 from a PaCOz of 35 mm Hg and 0.2 RI units subtracted for every 15 mm Hg increase in PaC02 above 35 mm Hg. It should be noted that there is minimal increase in Pa02 once a patient is on the RI 3 isobar. Here, if the patient is on 60 per cent 02 and the Pa02 is 90 mm Hg,
257
Goldfarb et al
placing the patient on 100 per cent 0s would give a Pa02 of 160 mm Hg. The amount of increase in the Pa02 decreases as the slopes of the isobars decrease. This reflects the generally accepted notion that if a patient is sufficiently ill to require 60 per cent Fro*, increasing the Fro:! beyond 60 per cent does not effect much of an increase in PaOz. In addition, the risk of oxygen toxicity is a hazard when the patient is placed on high concentrations for long periods [2]. Conclusions
There are several areas of research that require further investigation with regard to the respiratory index. They include: 1. Testing the RI with data from trauma patients in a prospective fashion. Subgroups of trauma patients may clarify the prognostic significance of the RI with regard to cases of isolated thoracic trauma, multiple trauma including the thorax, and multiple trauma excluding the thorax. The relevance of the RI in pulmonary units (nonsurgical) should prove helpful in tracking therapy. 2. The applicability of the nomogram will be tested further by the physicians handling respiratory problems as a graphic representation of the influence of various forms of therapy on respiratory pathophysiology. Various therapeutic regimens can be compared with similar groups of patients. 3. The prognostic and therapeutic value of the RI must be compared with more complex pulmonary indexes such as per cent of pulmonary shunting, pulmonary compliance, and functional residual capacity. Such indexes may perhaps be added to the RI so that respiratory pathophysiology can be characterized and collated even more precisely. The RI alone, however, provides a simple way to record data in a form that is easily understood.
258
Summary
The respiratory index (RI), P(AaDOz)/PaOz, was investigated in a retrospective study of 1’77 intubated patients treated at the Maryland Institute for Emergency Medicine. An RI of 0.1 to 0.37 is normal. Patients with an RI of 2 or greater were intubated. Those patients who reached an RI of 6 or more had an associated 12 per cent probability of survival. The RI reflects the presence of pulmonary shunting in a variety of circumstances including atelectasis, pulmonary contusion, and pulmonary emboli. A nomogram that allows the course of the patient with respiratory problems to be followed is described. Movement along the same isobars or between isobars can be followed by plotting the Pa02 against the Fro,. Thus, the rationale and effect of respiratory therapeutic manipulations may be recorded graphically. Acknowledgment: We wish to thank Mr Paul H. Broome for computerization of the data given herein and MS Marion Royston for technical assistance in the preparation of this report; both are from the Biomedical Laboratory, Edgewood Arsenal, Aberdeen Proving Ground, Maryland. References Nunn JF: Applied Respiratory Physiology. London, Butterworths. 1969, p 337. Pontoppidan H, Geffin B, Lowenstein E: Acute Respiratory Failure in the Adult. N Eng/ J Med 267: 743, 1972. Siegel JH: Personal communication. Siegel JH, Farrell EJ: A computer simulation model to study the clinical observability of ventilation and perfusion abnormalities in human shock states. Surgery 73: 898, 1973. 5. Laver MB, Austen WG: Cardiorespiratory dynamics, chapt 1, p 2. Surgery, 2nd ed. Philadelphia, Saunders, 1969. 6. Mellemgaard K: The alveolar-arterial oxygen difference: its size and components in normal man. Acta Pbysb/ Stand 73: 10, 1966.
The American Journal ol Surgery
