Abstract

Introduction: PaO2/FiO2 ratio (PF Ratio) is used in severity and prognostic scores, as well as clinically. Surrogates for the PF Ratio have also been validated against it; however, the PF Ratio has not been validated as a surrogate of gas-exchange status.

Aim of the Study: To provide evidence of the disconnect between the PF Ratio and the physiologic indicators of gas-exchange, and to suggest alternatives.

Materials and Methods: The weakness of the relationship between the PF Ratio and the A-a Gradient is demonstrated mathematically. New surrogates were tested with parameters obtained from patients with hypoxemia associated with COVID-19 infection. A “final” surrogate was chosen based on ease of calculation and familiarity with interpretation.

Results: Hypothetical-values for PaO2 and for FiO2 were paired, and PF Ratios and A-a Gradients were calculated accordingly. The equation for calculating the A-a Gradient was simplified to obtain surrogates for the A-a Gradient. A total of 264 sets of ABGs were used to calculate A-a Gradients. The algebraic difference (SpO2 – FiO2%), used interchangeably with (SaO2 – FiO2% + 3), showed very high correlation with the A-a Gradient (R² = 0.9685 and R² = 0.9457 respectively).

Conclusions: (SpO2 – FiO2%) and (SaO2 – FiO2% + 3) have a very high correlation with the A-a Gradient. The PF Ratios 100, 200, 300, and 400, can be replaced with 30, 55, 65, and 70 respectively. Finally, SpO2 of 93% or higher should be targeted to ensure SaO2 of “around” 90%.

Keywords: PF Ratio, PaO2, FiO2, A-a Gradient, SpO2, SaO2, ARDS, SOFA.

Introduction

The ratio of the Arterial Oxygen-Partial-Pressure (PaO2) to the Fractional-Concentration of Inspired-Oxygen (FiO2), also known as the PaO2/FiO2 or PF Ratio, has been in utility since at least1989, when it was proposed as an “early test of survival” in the Adult Respiratory Distress Syndrome (ARDS)¹ . With the significant increase in the number of patients with ARDS-like lung-disease, associated with the COVID-19 pandemic, there has been an increase in the use of the PF Ratio, not only as an indicator of severity of gas-exchange impairment, but also to monitor its progression over time. In my personal experience I have encountered situations when patients’ family-members asked specifically about the PF Ratio. One of the earliest scales that incorporated the PF Ratio among its criteria was the “Berlin Definition” and severity-score of ARDS ². Later on, the PF Ratio made its way into treatment protocols, like the NIH NHLBI ARDS Clinical Network (ARDSnet) Mechanical Ventilation Protocol ³, as well as other disease- severity scores like the Sepsis-related Organ Failure Assessment (SOFA) score ⁴. And, to be able to generate such data “non-invasively”, since PaO2 requires Arterial Blood Gases (ABGs), modifications of such scores were conceived using Oxygen-saturation obtained through Pulse- Oximetry (SpO2) ⁵. So, SpO2/FiO2 ratio thresholds, with or without qualification by Positive End- Expiratory Pressure (PEEP), were validated as surrogates for the PF Ratio thresholds ^{6, 7}. From there on the SpO2/FiO2 ratio took a life of its own, and it became the basis for parameters like the SpO2/FiO2 Time-at-Risk (SF-TAR) ⁸. However, except indirectly, as in association with mortality, the PF Ratio itself has never been validated against known physiologic parameters that are indicative of the gas-exchange status^{1, 9}. Even when validated in association with mortality, it was noticed that supplementing the PF Ratio with the FiO2 improves its association with mortality. In an analysis of data from the ARDSnet studies¹⁰, in which the primary purpose was “to explore the potential value of adding PEEP and FiO2 criteria to the AECC criteria to exclude patients with low risk of mortality”, the authors concluded that “the addition of baseline PEEP would not have increased the value of PaO2/FiO2 for predicting mortality of ALI/ARDS patients.

In contrast, the addition of baseline FiO2 to the PaO2/FiO2 Ratio could be used to identify subsets of patients with low or high mortality.” Requiring both, the PF Ratio and FiO2, means that both PaO2 and FiO2 will be known, and that their ratio is part of an “equation” rather than being an independent parameter.

Ventilation-based parameters, related to PaCO2, minute-ventilation, and fraction of dead-space (Vd/Vt) have also been proposed for prognostic as well as monitoring purposes^{11, 12}; however, calculating those parameters is not simple, and they are unlikely to be retrofitted into established scales and scores that involve the PF Ratio.

PaO2/FiO2 may sound like a reasonable “efficiency” index: the amount of Oxygen put into the circulation relative to the concentration of inspired-Oxygen. However, PaO2 does not represent an amount of blood extracted from the alveolar-space; it represents the partial-pressure ofOxygen in the blood at which the flow of Oxygen, from the alveolar-space to the circulation, will stop. So, regardless of how low or how high the Mixed-Venous Oxygen Partial-Pressure (PvO2) is, PaO2 will be the same, granted that enough time is allowed for the blood to pass through the lungs. This is why, for example, a Left-to-Right shunt does not result in higher PaO2, even though PvO2 is higher than usual.

The flow of Oxygen across the Alveolar-capillary interface stops because the resistance to flow (or friction, drag, …) becomes equivalent to the driving-force between the two sides. This driving force is the “algebraic – difference” between the Oxygen partial-pressures on either side of the Alveolar-capillary interface: the Alveolar-arterial (partial-pressure) gradient, better known as the A-a Gradient. This gradient is the variable that changes with disease and recovery, and consequently, causes the arterial Oxygen Partial-Pressure (PaO2) to change if the Alveolar Oxygen Partial-Pressure (PAO2) is unchanged (for example, PaO2 on Room-Air will decrease because the A-a Gradient becomes wider, or steeper).

 

Aim of the Study

This work is intended to investigate the correlation, or lack thereof, between the PF Ratio and the A-a Gradient. It is also intended to provide alternative surrogates, which have better correlation, and to enable the use of the new surrogates in scales and scores that included the PF Ratio among their parameters.

 

Materials and Methods

Mathematical modeling was used to provide representation of the relationship between the PF Ratio and the A-a Gradient, using theoretical pairs for PaO2 and FiO2, and commonly-used default-values for the remaining variables in the A-a Gradient equation. Subsequently, the A-a Gradient equation was simplified, using approximation and elimination, to derive alternative surrogates. The newly-derived surrogates were then tested against A-a Gradients calculated from 264 blood-gases obtained from 10 patients suffering from acute hypoxemic respiratory failure associated with COVID-19 infection. A final “best” surrogate for the A-a Gradient was chosen based on ease of calculation and on familiarity with interpretation. Finally, values for the new surrogate were chosen in order to be implemented and retrofitted into scales and scores that incorporate the PF Ratio as one of their parameters. For those new values to have the best representation of the PF Ratio, they were determined manually from the real patient data, and then extrapolated to include values that were not represented in the current sample.

 

Results

Hypothetical-values for PaO2 (natural numbers from 40 to 200), and for FiO2 (in percentage form or FiO2%, multiples of 5 from 30 to 100) were paired. A-a Gradients were calculated using Patm 760 mmHg, PH2O 47 mmHg, PCO2 40 mmHg, and RQ 0.8. The resultant graph is depicted in Figure-1.It should be obvious from Figure-1 that, with unchanged gas-exchange, hence constant A-a Gradient, the PF Ratio can change dramatically depending on FiO2. Every value for PF Ratio corresponds to a range of A-a Gradient values, and vice-versa. Most importantly, when the PF Ratio changes, up or down, there is no way of knowing in which direction the A-a Gradient is changing, if changing at all.

From mathematical derivations that I had performed (see Appendix A) the following expressions were determined to be surrogates for the A-a Gradient:

1- (7 x FiO2%) – PaO2

2- (2 x FiO2%) – SaO2

3- (FiO2% – SaO2)

In order to test the above-mentioned theoretical surrogates, a total of 264 Arterial-Blood-Gas measurements from 10 critically-ill COVID-19 patients were used. The parameters had been corrected to patient-temperature. In 254 of the blood-gas sets the Oxyhemoglobin (SaO2) was measured; blood-gas sets without measured SaO2 were not used in calculations involving Oxygen saturation. Simultaneous peripheral SpO2 values were available for 181 of the blood-gas sets. Both measured SaO2 and SpO2 were available in 172 sets, and only one set had neither a measured SaO2 nor an SpO2.

Except for one patient on a non-rebreather mask and 9 patients on high-Flow heated-humidified Oxygen, all the other blood-gas sets were obtained when the patients were on mechanical ventilation. And, except for one set obtained on PEEP of 5 cm H2O and another set on PEEP 20 cm H2O, PEEP ranged between 6 and 18 cm H2O.

The following regression-equations were derived from the 264 sets of ABGs (see Appendix B).

A-a Gradient = 7.01(FiO2%) – 1.06(PaO2) – 50.87

A-a Gradient = 6.80(FiO2%) – 5.17(SaO2) + 352.91

A-a Gradient = 6.78(FiO2%) – 3.45(SpO2) + 208.81

(SpO2 – FiO2%) = 0.9982(SaO2 – FiO2%) + 3.261

(SpO2 – FiO2%) and [(SaO2 – FiO2%) + 3] may be used interchangeably. A-a Gradient was plotted as a function of (SpO2 – FiO2%) and (SaO2 – FiO2% + 3), and the resulting graph is depicted in Figure-2.

In order to “retrofit” the new surrogates into scales and scores that included the PF Ratio among their parameters, thresholds had to be chosen for the newly generated surrogates. By definition, there are no “exact” equivalents for the PF Ratio thresholds; however, the “best” corresponding-thresholds would be those that would make the results most accurate: include the most values in the new blocks that would have been included in the corresponding old blocks, and to exclude the most values from the new blocks that would have been excluded from the old blocks.

For example, if A, B, C, and D corresponded to 100, 200, 300, and 400 respectively, then the values between A and B should include the largest number of PF Ratios between 100 and 200, and the least number of PF Ratios that are not. I performed a manual search for such thresholds (see Appendix C), but since the PF Ratios in my 264 ABGs ranged between 39 and 383, I wasn’t able to obtain a threshold that corresponds to a PF Ratio of 400. The values obtained were A-a Gradients of 370 mmHg, 155 mmHg, and 92 mmHg for PF Ratios of 100, 200, and 300, respectively.

The best fit for the three new A-a Gradient-thresholds against their corresponding PF Ratio- Thresholds was a “power-function”, with the following regression-equation:

A-a Gradient = 125933 x (PF^–1.266)

(R² = 1)

Obviously, three numbers hardly make a “trend”; however, the best fit for the theoretical PF Ratios against A-a Gradient, for PaO2 values between 50 and 80 mmHg, was:

A-a Gradient = 129600 x (PF^–1.297)

(R² = 0.9036)

The almost-identical regression-equations are a mere coincidence, and as can be seen from the regression-equation generated from all 264 points (Figure-3), the closeness of the equations is restricted to a narrow, albeit clinically-pertinent, range of PaO2 values.

Table-2 lists values for (SpO2 – FiO2%) and (SaO2 – FiO2% + 3) that correspond to PF Ratio- thresholds used in prognostic scales.

The theoretical model [A-a Gradient = 129600 x (PF^–1.297)] was used to derive the A-a Gradients, and the remaining values were derived from their corresponding regression-equations. The “115” and “175” PF Ratios were added because they were used in the reanalysis of the AECC data for the purpose of evaluating the contribution of adding PEEP and/or FiO2 to the PF Ratio¹⁰.

 

Figure-1 PF Ratio and A-a Gradient (theoretical)

Figure-2 Alternatives for PF Ratio

Figure-3 Trend-lines for PF Ratio and A-a Gradient

Table 1

Table 2

Please click here to view all figures and tables

 

Discussion and Conclusions

(SpO2 – FiO2%) interchangeably with (SaO2 – FiO2% + 3), should replace the PaO2/ FiO2 Ratio for a more physiologic, and consistent, surrogate for gas-exchange abnormalities (A-a Gradient).

Multiples-of-five for PF Ratios of 100, 115, 175, 200, 300, and 400 would be 30, 35, 55, 60, 65,and 70 respectively. The new/ alternative thresholds for the PF Ratio-thresholds can be retrofitted into clinical and research scores into which the PF Ratio has been incorporated.

From the above discussed equations, and from previous reports¹³ on the difference between SpO2 and SaO2, SpO2 of 93% or higher should be targeted to ensure an SaO2 of “around” 90%. Also, the regression-equations and the high correlation of the theoretically-derived surrogates indicate that a change of 5 percentage points in FiO2% can be expected to cause a change of 5-10 percentage points in SpO2; so, consideration should be given to using increments (and decrements) of 2-3 percentage in FiO2%, instead of the customary 5 percentage points, in order to avoid major swings in SpO2%, up and down, which would require frequent re-adjustments.

 

References  

1. Bone RC, Maunder R, Slotman G, Silverman H, Hyers TM, Kerstein MD, Ursprung JJ. An early test of survival in patients with the adult respiratory distress syndrome. The PaO2/FiO2 ratio and its differential response to conventional therapy. Prostaglandin E1 Study Group. Chest. 1989 Oct;96(4):849-51.

2. Bernard GR, Artigas A, Brigham KL, Carlet J, Falke K, Hudson L, 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(3):818-24.

3. (n.d.). http://www.ardsnet.org/files/ventilator_protocol_2008-07.pdf

4. Vincent JL, Moreno R, Takala J, Willatts S, De Mendonça A, Bruining H, et al. The

SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine. Intensive Care Med. 1996;22(7):707-10.

5. Namendys-Silva SA, Silva-Medina MA, Vásquez-Barahona GM, Baltazar-Torres JA, Rivero-Sigarroa E, Fonseca-Lazcano JA, et al. Application of a modified sequential organ failure assessment score to critically ill patients. Braz. J. Med. Biol. Res. 2013;46(2):186-93.

6. Rice TW, Wheeler AP, Bernard GR, Hayden DL, Schoenfeld DA, Ware LB; National Institutes of Health, National Heart, Lung, and Blood Institute ARDS Network. Comparison of the SpO2/FiO2 ratio and the PaO2/FiO2 ratio in patients with acute lung injury or ARDS. Chest. 2007 Aug;132(2):410-7.

7. Pandharipande PP, Shintani AK, Hagerman HE, et al. Derivation and validation of SpO2/FiO2 ratio to impute for PaO2/FiO2 ratio in the respiratory component of the Sequential Organ Failure Assessment score. Crit Care Med. 2009;37(4):1317-1321.8. Adams JY, Rogers AJ, Schuler A, Marelich GP, Fresco JM, Taylor SL, et al. Association Between Peripheral Blood Oxygen Saturation (SpO2)/Fraction of Inspired Oxygen (FiO2) Ratio Time at Risk and Hospital Mortality in Mechanically Ventilated Patients. Perm J. 2020;24.

9. ARDS Definition Task Force, Ranieri VM, Rubenfeld GD, Thompson BT, Ferguson ND, Caldwell E, Fan E, Camporota L, Slutsky AS. Acute respiratory distress syndrome: the Berlin Definition. JAMA. 2012 Jun 20;307(23):2526-33.

10. Britos M, Smoot E, Liu KD, Thompson BT, Checkley W, Brower RG, et al. The value of positive end-expiratory pressure and FiO? criteria in the definition of the acute respiratory distress syndrome. Critical care medicine. 2011;39(9):2025-30.

11. Nuckton TJ, Alonso JA, Kallet RH, Daniel BM, Pittet JF, Eisner MD, Matthay MA. Pulmonary dead-space fraction as a risk factor for death in the acute respiratory distress syndrome. N Engl J Med. 2002;346:1281–1286.

12. Sinha, P, Calfee, CS, Beitler, JR, Soni, N, Ho, K, Matthay, MA, et al. Physiologic Analysis and Clinical Performance of the Ventilatory Ratio in Acute Respiratory Distress Syndrome. Am. J. Respir. Crit. Care Med. 199:3, 333-341.

13. Van de Louw A, Cracco C, Cerf C, Harf A, Duvaldestin P, Lemaire F, Brochard L. Accuracy of pulse oximetry in the intensive care unit. Intensive Care Med. 2001 Oct;27(10):1606-13..\

 

Appendix A

 The original A-a Gradient equation is as follows:

 A-a Gradient = [(P_atm – P_H2O) x FiO₂] – (PCO₂/RQ) – PaO₂

Since the contribution of PaCO₂/RQ will be to the intercept, by eliminating it we reach a parameter that is an approximation of the A-a Gradient, albeit predictably smaller by around 50 mmHg; so,

 [(P_atm – P_H2O) x FiO₂] – (PCO₂/RQ) – PaO₂

 becomes

 (713 x FiO₂) – PaO₂.

 More often than not, in clinical practice, we communicate FiO₂ not as a decimal fraction (0.21, 0.50, etc…) but as a percentage (21%, 50%, etc…). Since FiO₂% = 100 x FiO₂, the new parameter ends up looking like this:

 (7.13 x FiO₂%) – PaO₂

So, simplifying the equation, we can have a surrogate for the A-a Gradient in the simple form of:

(7 x FiO₂%) – PaO₂

Figure-A-1 depicts the correlation between [(7 x FiO₂%) – PaO₂] and the A-a Gradient.

Figure-A-2 depicts the Oxygen-Dissociation Curve; the part that pertains to our clinical practice in the Intensive Care Unit (ICU) is the one corresponding to SaO₂ between 90% and 100%. These values correspond, roughly, to PaO₂ from 60 to 100 mmHg, respectively.

That part of the curve in which we would be interested is almost linear, and can be approximated by the following equation:

SaO₂ ≈ (PaO₂/4) + 75.

Incorporating this equation into the calculation of the A-a Gradient, we get the following:

A-a Gradient ≈ (7.13 x FiO₂%) – 50 – 4(SaO₂ – 75) ≈ (7.13 x FiO₂%) – (4 x SaO₂) + 250.

Again, since we are looking for a surrogate and not, necessarily, an approximation, dropping the “+ 250” will only affect the intercept. Also, rounding the 7.13 up to 8 makes for an easier equation to handle and simplify:

(8 x FiO₂%) – (4 x SaO₂).

Since dividing by 4 will make all the answers 4 times smaller, the correlation between A-a Gradient and the surrogate will be maintained. So, the following “expression” can be a surrogate of the A-a Gradient:

(2 x FiO₂%) – SaO₂

Figure-A-3 depicts the relation between [(2 x FiO₂%) – SaO₂] and the A-a Gradient.

The close correlations, in Figure-A-1 and Figure-A-3, are not a surprise, since there were only minor modifications to the original equation. However, even a very simple, and barely physiologic, expression, (FiO₂% – SaO₂) has a much better correlation, with the A-a Gradient, and much less overlap, than the PF Ratio (Figure-A-4).

A-a Gradient as a function of FiO₂%, without “fine-tuning” the latter with PaO₂ or SaO₂, is depicted in Figure-A-5. The “short” portions within each column of points represent the values of FiO₂% and their corresponding A-a Gradients when PaO₂ is between 40 and 80 mmHg. Among the latter values, there is minimal, if any, overlap; however, every FiO₂% still represents a range of A-a Gradients (30 mmHg, which is the range of PaO₂ from 50 to 80 mmHg).

 

Figure-A-1 (7FiO2 - PaO2) and A-a Gradient (theoretical)

Figure-A-2 Oxygen-Dissociation Curve

Figure-A-3 (2FiO2 - SaO2) and A-a Gradient (theoretical)

Figure-A-4 (FiO2 - SaO2) and A-a Gradient (theoretical)

Figure-A-5 FiO2 and A-a Gradient (theoretical)

 

Appendix B

Using the statistical-package in Microsoft Excel 2010 and the same default values for P_atm and RQ, a regression-equation was generated for A-a Gradient as a function of FiO₂% and PaO₂; the coefficients were as follows:

R² = 0.99

Intercept = –50.87 (P-value 4.88E–28)

FiO₂% coefficient = 7.01 (P-value 7.58E–252)

PaO₂ coefficient = –1.06 (P-value 2.63E–88)

So, the regression-equation for A-a gradient as a function of and FiO₂% and PaO₂ would be:

A-a Gradient = 7.01(FiO₂%) – 1.06(PaO₂) – 50.87

This confirms the theoretical approximation.

Figure-B-1 depicts the A-a Gradient as a function of [(7 x FiO₂%) – PaO₂] using values from the 264 sets of ABGs (and same default values for P_atm and RQ).

Using the same statistical-package in Microsoft Excel 2010 and the same default values for P_atm and RQ, a regression-equation was generated for A-a Gradient as a function of FiO₂% and SaO₂; the coefficients were as follows:

R² = 0.97

Intercept = 352.91 (P-value 2.43E–24)

FiO₂% coefficient = 6.80 (P-value 8.58E–196)

SaO₂ coefficient = –5.17 (P-value 9.52E–39)

So, the regression-equation for A-a gradient as a function of and FiO₂% and SaO₂ would be:

A-a Gradient = 6.80(FiO₂%) – 5.17(SaO₂) +352.91

So, the regression-equation for A-a Gradient as a function of FiO₂% and SaO₂ is not very close to the theoretically-derived (2FiO₂% – SaO₂), but not very far off from (FiO₂% – SaO₂).

Using the same statistical-package in Microsoft Excel 2010 and the same default values for P_atm and RQ, a regression-equation was generated for A-a Gradient as a function of FiO₂% and SpO₂; the coefficients were as follows:

R² = 0.96

Intercept = 208.81 (P-value 6.51E–4)

FiO₂% coefficient = 6.78 (P-value 8.72E–117)

SpO₂ coefficient = –3.45 (P-value 4.12E–8)

So, the regression-equation for A-a gradient as a function of and FiO₂% and SpO₂ would be:

A-a Gradient = 6.78(FiO₂%) – 3.45(SpO₂) + 208.81

The ratio of the FiO₂% coefficient to the SpO₂ coefficient (–1.97:1) is much closer to the theoretically-derived (–2:1) than the ratio in the SaO₂ regression-equation (–1.3:1).

(FiO₂% – SaO₂) and (FiO₂% – SpO₂) would be expected to generate negative values in most situations, and they move in the opposite direction to the changes in gas-exchange; reversing the order of the variables eliminates those concerns.

Figure-B-2 depicts the relation between (SpO₂ – FiO₂%) and (SaO₂ – FiO₂%).

The regression-equation for (SpO₂ – FiO₂%) as a function of (SaO₂ – FiO₂%) would be:

(SpO₂ – FiO₂%) = 0.9982 (SaO₂ – FiO₂%) + 3.261.

 

Figure-B-1 (7FiO2% - PaO2) and A-a Gradient (264 ABGs)

Figure-B-2 (SaO2 - FiO2%) and (SpO2 - FiO2%) (172 ABGs)

 

Appendix C

A-a Gradient (A-a G)	PF Ratio (PF)		Total Prevalence	Sensitivity	Specificity	Accuracy
	PF <100	(Not) <100		0.73	0.92	0.86
(A-a G) >370	61	23	84
(Not) >370	15	165	180
Total Test Results	76	188	264

Table-C-1: Correct ID of ABGs with [(PF Ratio) < 200] by [(A-a Gradient) > 370]

A-a Gradient (A-a G)	PF Ratio (PF)		Total Prevalence	Sensitivity	Specificity	Accuracy
	100≤ PF <200	100≤ (Not) <200		0.79	0.73	0.77
155< (A-a G) ≤370	124	32	156
155< (Not) ≤370	29	79	108
Total Test Results	153	111	264

Table-C-2: Correct ID of ABGs with [100≤ (PF Ratio) <200] by [155 < (A-a Gradient) ≤ 370]

 

A-a Gradient (A-a G)	PF Ratio (PF)		Total Prevalence	Sensitivity	Specificity	Accuracy
	200≤ PF <300	200≤ (Not) <300		0.77	0.93	0.92
92< (A-a G) ≤155	17	5	22
92< (Not) ≤155	16	226	242
Total Test Results	33	231	264

Table-C-3: Correct ID of ABGs with [200≤ (PF Ratio) <300] by [92 < (A-a Gradient) ≤ 155]

 

A-a Gradient (A-a G)	PF Ratio (PF)		Total Prevalence	Sensitivity	Specificity	Accuracy
	PF ≥300	(Not) ≥300		0.50	1.00	0.99
(A-a G) ≤92	1	1	2
(Not) ≤92	1	261	262
Total Test Results	2	262	264

Table-C-4: Correct ID of ABGs with [(PF Ratio) ≥300] by [(A-a Gradient) ≤ 92]

A-a Gradient (A-a G)	PF Ratio (PF)		Total Prevalence	Sensitivity	Specificity	Accuracy
	Correct Block	Not		0.77	0.77	0.92
Total within Blocks	203	61	264
Total outside Blocks	61	731	792
Total Test Results	264	792	1056

Table-C-5: Correct ID of ABGs with (PF Ratio) in correct blocks by (A-a Gradients) 92, 155, and 370.