Using a rapidly responding nitric oxide (NO) analyzer, we measured the steady-state NO diffusing capacity (Dl NO) from end-tidal NO. The diffusing capacity of the alveolar capillary membrane and pulmonary capillary blood volume were calculated from the steady-state diffusing capacity for CO (measured simultaneously) and the specific transfer conductance of blood per milliliter for NO and for CO. Nine men were studied bicycling at an average O2 consumption of 1.3 ± 0.2 l/min (mean ± SD). Dl NO was 202.7 ± 71.2 ml · min−1 · Torr−1 and steady-state diffusing capacity for CO, calculated from end-tidal (assumed alveolar) CO2, mixed expired CO2, and mixed expired CO, was 46.9 ± 12.8 ml · min−1 · Torr−1. NO dead space = (Vt × Fe NO − Vt × Fa NO)/(Fi NO − Fa NO) = 209 ± 88 ml, where Vt is tidal volume and Fe NO, Fi NO, and Fa NO are mixed exhaled, inhaled, and alveolar NO concentrations, respectively. We used the Bohr equation to estimate CO2 dead space from mixed exhaled and end-tidal (assumed alveolar) CO2 = 430 ± 136 ml. Predicted anatomic dead space = 199 ± 22 ml. Membrane diffusing capacity was 333 and 166 ml · min−1 · Torr−1 for NO and CO, respectively, and pulmonary capillary blood volume was 140 ml. Inhalation of repeated breaths of NO over 80 s did not alter Dl NO at the concentrations used.
- alveolar capillary gas diffusion
- dead space
- membrane diffusing capacity
- lung capillary blood volume
theoretically,nitric oxide (NO) is the ideal gas to study lung diffusion because it is poorly soluble in water and, because of its very rapid and virtually irreversible reaction with Hb, its uptake is independent of pulmonary capillary blood flow and rate of chemical reaction (6). Our laboratory (4) originally developed a single-breath method for measuring lung NO diffusion (Dl NOsb) during breath holding at rest. By including CO in the inspired mixture, it is possible to simultaneously measure the Dl COsb. From knowledge of the water solubility and molecular weight of the two gases and by assuming that Roughton and Forster's model also applies to NO lung transfer, our laboratory (3) and others (8) have estimated the Dm and Vc from a single breath. The availability of rapidly responding and sensitive NO analyzers now allows estimation of alveolar NO concentrations from end-expiratory samples and hence measurement of steady-state Dl NO(Dl NOss) during tidal breathing using safe and stable NO concentrations.
These measurements are important for several reasons. NO is produced in high concentration (11) in the nose and taken up in the lung, so steady-state NO gas transfer takes place during everyday life and knowledge of Dl NO is essential for understanding normal NO metabolism. Measuring diffusing capacity in animals and on maximal exercise in humans is easier using a steady-state method compared with a single-breath technique. Using Roughton and Forster's method for obtaining Dm and Vc from Dl CO during steady-state breathing at two or more O2 tensions is likely to overestimate the O2-dependent resistance (1/θco Vc) due to hyperoxia increasing regional inhomogeneity in diffusing capacity. Using Dl NO and Dl CO at a single, physiological value for PcO2 obviates this problem. Exhaled NO is being investigated as a test for inflammatory markers in lung disease. An increase in Fe NO could result from increased NO production or from reduced Dl NO(10). Ill or ventilated patients will be unable to perform single exhalations from total lung capacity, so knowing Dl NOss is important in interpreting values for Fe NO derived from tidal breathing measurements. Finally, inhaled NO is being used therapeutically to treat acute respiratory distress syndrome and pulmonary hypertension. NO causes ultrastructural oxidant lung injury when inhaled in concentrations of 6 ppm for 6 wk (12), and its safety depends on its fast removal into the pulmonary capillary blood before oxidation to nitrogen dioxide or other toxic products can occur. If patients have reduced Dl NOss, then toxicity may be enhanced and the inhaled NO concentration may need to be reduced.
For these reasons, we have made combined Dl NOss and Dl COss measurements in healthy volunteers by adapting the method of Bates et al. (1).
- Nitric oxide [parts per billion (ppb)]
- Lung diffusing capacity
- Diffusing capacity for NO
- Diffusing capacity for CO
- Steady-state diffusing capacity for CO (ml · min−1 · Torr−1)
- Steady-state diffusing capacity for nitric oxide (ml · min−1 · Torr−1)
- Single-breath diffusing capacity for CO (ml · min−1 · Torr−1)
- Single-breath diffusing capacity for NO (ml · min−1 · Torr−1)
- Tidal volume (liters)
- Alveolar volume (liters)
- Respiratory rate (min−1)
- Barometric pressure (Torr)
- Specific transfer conductance of blood per milliliter for NO (ml · min−1 · Torr−1 · ml−1)
- Specific transfer conductance of blood per milliliter for CO (ml · min−1 · Torr−1 · ml−1)
- Alveolar CO2 concentration (%)
- Mixed exhaled CO2 concentration (%)
- Mixed exhaled CO concentration (%)
- Inhaled CO concentration (%)
- Inhaled NO concentration (ppb)
- Mixed exhaled NO concentration (ppb)
- Alveolar NO concentration (ppb)
- Dead space (ml)
- Diffusing capacity of alveolar capillary membrane (ml · min−1 · Torr−1)
- Pulmonary capillary blood volume (ml)
- Partial pressure of O2 in pulmonary capillaries (Torr)
- O2 uptake (l/min)
- Concentration of hemoglobin in venous blood (g/dl)
- Concentration of carboxyhemoglobin venous blood (%)
- Concentration of methemoglobin in venous blood (%)
Nine nonsmoking male subjects with no history of lung disease volunteered for the study. All signed consent forms approved by the Huntingdon district ethics committee, who also gave approval for this study. Their characteristics and Dl COsb measured by the standard technique at rest (6) are listed in Table1. All subjects pursued a moderately active lifestyle.
On a daily basis, temperature and Pb were measured by use of a metabolic cart (2900Z, Sensor Medics EME, Brighton, Sussex, England) calibrated against a mercury thermometer and the local meteorological office, respectively. Pb was corrected for saturated vapor pressure by reference to a table (6). Each subject performed one complete maneuver. Background atmospheric NO, CO2, and CO were measured. Each subject sat breathing air on an electronically braked bicycle ergometer (Sensor Medics 800S) and started to cycle, gradually building up speed until he achieved the desired level of moderate exercise (∼1 l/minV˙o 2). Once steady state was achieved, exhaled air was collected for a 2-min period to measure NO and CO back tension, and an 80-s record was also taken of intrabreath NO and CO2 recording at the lips by arranging a fine-bore cannula within the mouthpiece to ensure zero dead space. They were then switched to a mixture of ∼5,000 ppb NO and 0.1% CO in air stored in a 200-liter Douglas bag (PK Morgan, Gillingham, Kent, UK) prepared immediately before each replicate. A continuous 2-min collection of exhaled air was made from a similar Douglas bag attached to the exhaled port of the metabolic cart. Vt on a breath-by-breath basis, RR, and V˙o 2 were recorded during this time from the metabolic cart. Volume was calibrated by using a 3-liter syringe. The O2 analyzer of the metabolic cart was calibrated using three different O2concentrations (16%, 21%, and 26%). Immediately before and immediately after the period of exercise, the inspirate bag was analyzed for NO and CO.
NO and CO2 were analyzed by using a rapidly responding instrument (Logan Research LR 2000, Rochester, UK) that directs the sample through a rapidly reacting infrared CO2analyzer to a chemiluminescent NO analyzer with a small (<10 ml) reaction chamber. The NO analyzer has four separate analysis modes for quantifying endogenous NO produced 1) through the nose, 2) during tidal breathing, and 3) during a maximal single exhalation and, finally, 4) a mode for analyzing inhaled concentrations during therapeutic use of NO. For this study, the instrument was set in “therapeutic” mode and calibrated using NO-free compressed air and 4,000 ppb and 80,000 ppb NO (manufacturer's certificate of analysis; BOC gases, Worsley, Manchester, UK). The CO2 analyzer was calibrated using O2 (CO2 free), 5% and 6% CO2. To ensure linearity of these two analyzers, a serial dilution of a mix of NO and CO2 over the working range was performed. To measure the response time, a rubber balloon containing 2,660 ppb NO and 2.7% CO2 was burst by pinprick.
CO was analyzed by using the infrared analyzer of a standard gas transfer apparatus (Transfertest, PK Morgan, Chatham, Kent, UK). Linearity was tested by serially diluting a standard CO-He mix of 6% He and 0.1% CO in air to generate a 25-point plot of CO vs. He.
Venous blood was sampled immediately before each experiment and analyzed within 2 h for Hb, metHb, and COHb using an automated spectrophotometer (IL282, Instrument Laboratories, Cupertino, CA).
The exercise laboratory was kept well ventilated at all times. The stock NO cylinder (1,000 ppm in nitrogen) was kept securely fastened to the purpose-built trolley at all times. The inspired NO concentration in the Douglas bag was checked before inhalation and was continuously monitored breath by breath during the experiment.
Alveolar NO and CO2.
The NO, CO2, and time (in s) readings were downloaded as Excel spreadsheet files (Microsoft) from the Logan analyzer using the “datadump facility.” For each individual, four files were created: mixed exhaled (breathing air on exercise), mixed exhaled (breathing NO and CO on exercise), inhaled (breathing NO and CO on exercise), and breath by breath (breathing NO and CO on exercise.) For the inhaled file, the concentrations were derived as the mean of the column of readings. For the breath-by-breath files, X-Y plots of NO as a function of time and CO2 as a function of time were drawn using a spreadsheet charting program (Works 3.1, Microsoft). The FaCO2 was taken as one-third of the way along the alveolar plateau (Fig. 1) (5). For NO, the alveolar phase was identified visually as a stepwise reduction in the rate of fall of NO concentration with time (Fig. 2), a line of best fit was drawn, and the concentration one-third of the way along was taken as Fa NO. Fe CO and Fe NO were taken as the mean of the column of readings from the exhaled bag.
Dead space calculations.
Diffusing capacity calculations.
We calculated Dl CO and Dl NO by using estimated anatomic dead space [(a) DCO I in Bates et al. (1)] and also Dl CO using CO2 dead space [(e) in Ref. 1] Equation 3 Equation 3 was calculated using the predicted anatomic dead space for both gases and, for CO, the CO2dead space (Eq. 1 ).
In addition, for NO, because we had a rapidly reacting NO analyzer, we adapted Bates et al.'s method using end-tidal concentrations [(c) DCO II in their paper] for NO and used end-tidal Fe NO from the expired NO curve (Fig. 2) as an estimate of Fa NO Equation 4There were thus two estimates of Dl COand Dl NO per subject.
For dead space and Dl calculations, the respiratory quotient was assumed to be 1 (see results).
Calculation of Dm and Vc.
Calculations were made as follows Equation 5 Equation 6 Equation 7For derivation of these equations, see Ref. 4. The value of θno (5) was taken as 4.5 ml · min−1 · Torr−1 · ml−1(1,500 mmol · min−1 · kPa−1 · l−1) and θco (7) calculated from 1/θco = 1.3 × 10−3 + 4.1 × 10−3 (PcO2) min · Torr. PcO2 was taken as 100 Torr. For both CO and NO, correction for Hb was made by multiplying θ by Hb × (1 − COHb − metHb)/14.6.
The paired t-test was used to compare the calculated NO dead space to the predicted anatomic dead space and to the CO2 dead space and to compare Dl NOcalculated from Fa NO and predicted anatomical dead space. Values are given as means ± SD.
Performance of Analyzers
The NO-CO2 analyzer was linear to serial dilution over the working range down to NO concentrations below 250 ppb and CO2 below 0.55% when departure from linearity occurred; these were well below the alveolar and mixed expired concentrations. The minimum detectable concentration of NO was 60 ppb and CO2 0.3% in therapeutic mode. The response time to half signal after balloon burst was 160 ms for CO2 and 400 ms for NO; the CO2 column was therefore “cut and pasted” down by 240 ms in the spreadsheet to synchronize concentration data for the two gases. A plot of CO and He was linear with intercept zero over the working CO range.
Table 1 illustrates the nine subjects' physical characteristics.
Measured and Derived Variables
Mean V˙o 2 was 1.3 ± 0.2 l/min. Mean respiratory quotient (RQ) was 0.96 ± 0.06. No loss of NO or CO occurred from the inspired bag. There was no detectable (i.e., <60 ppb) background NO in the laboratory atmosphere or in the exhaled breath sample for any subject with the instrument in the therapeutic mode. The breath-by-breath X-Y plots of exhaled concentration and time are shown in Fig. 1 for CO2 and Fig.2 for NO. The NO plot comprises a peak of inhaled NO, a steep fall representing dead space, and then a shallower alveolar slope. The mean inhaled, mean mixed exhaled, and alveolar NO were 5,218 ± 1,675, 1,402 ± 381, and 942 ± 250 ppb, respectively.
The calculated CO2 and NO dead spaces are shown in Table2. The NO dead space is significantly different (P < 0.001) from the CO2 dead space but not from the anatomic dead space. Calculated Dl CO and Dl NO together with Dm and Vc are shown in Tables 2 and 3. There are no significant differences between the two estimates of Dl NO. Dl NO appears constant over 30 s of observation (Fig. 3).
Estimates of dead space.
The NO dead space is strikingly similar to the predicted anatomical dead space but significantly lower, on average about one-half, than the CO2 dead space. However, the estimate of NO dead space is less precise because the alveolar slope for NO is steeper. Clearly, NO uptake is occurring within the alveolar dead space. NO is a reactive gas particularly in solution, and this reaction could be with groups other than Hb in the lining of unperfused alveoli. Heller and Schuster (9), however, discounted a chemical reaction of NO with lung tissue during single-breath Dl NO measurements in rabbits because there was no variation in Dl NO either with duration of breath hold or with Fi NO. Their results are in agreement with our single-breath observations of a semilog decline in Fa NO with intercept 1 in two human subjects (4) and identical values of Dl NOsb at Fi NO of 0.75 ppm and 60 ppm (Borland C and Cox Y, unpublished observations). A more plausible explanation is as follows: the physiological dead space exceeds the anatomic dead space because some alveoli have a ventilation-perfusion ratio approaching infinity. Such alveoli and their associated capillaries will secrete minimal CO2. On the other hand, Dl NO is diffusion limited, independent of capillary blood flow, and (if θno approaches infinity) independent of capillary blood volume. Uptake of NO would therefore occur even in these alveoli, and the NO dead space would correspond to the anatomic dead space.
Values for DlNO.
The values for Dl NO using the predicted anatomic dead space and the dead space calculated from Fa NO are similar (Table 2). The higher mean value for Dl NO calculated from Fa NO is largely due to subject 3, who had wide between-breath variation in Vt and Fa NO. The average Dl NOss of ∼200 ml · min−1 · Torr−1 measured here on moderate exercise exceeds the single-breath value measured at rest (125 ml · min−1 · Torr−1) by ∼70 ml · min−1 · Torr−1(4). This is at first surprising because steady-state measurements are made at Vt rather than total lung capacity and Dl NOsb is highly lung volume dependent (4), falling by 45 ml · min−1 · Torr−1 with a reduction in Va from 7 to 4 liters. It is to be noted that it is the volume that the inhaled NO is distributed into rather than the inhaled volume that is important here. During steady-state breathing, the NO will be distributed into a volume of functional residual capacity plus Vt, whereas during a single breath it will be distributed into Va. The difference between these volumes is perhaps only 1–2 liters. However, we also noted a rise in Dl NOsbwith exercise to 210 ml · min−1 · Torr−1 at 1–2 l/min V˙o 2, so the overall value could be little changed. These average values therefore may be used to interpret values for alveolar partial pressure of NO in models of NO production (10).
Values for DlCO, Dm, and Vc.
The mean Dl COss on moderate exercise exceeds Dl COsb measured at rest by ∼5–10 ml · min−1 · Torr−1. Our previous mean value for Dl COsb at 1.4 l/min V˙o 2 in three individuals was 48 ml · min−1 · Torr−1, so again it is possible that the discrepancy is due to exercise. Other workers have found Dl COss to be less than Dl COsb (13). DmCOss is close to our previous value for DmCOsb (316 ml · min−1 · Torr−1). However, Vc is about three times as large (56 ml single breath) (3). It is possible to use Eq. 7 to calculate Vc on exercise from our previous combined single-breath data (4), and a figure of 101 ml is obtained. In contrast, DmCOsb is 343 ml · min−1 · Torr−1. It would appear that during tidal breathing on exercise Vc is one to two times that which is measured during apnea at total lung capacity on exercise. One explanation is that the capillaries are flattened when the alveoli are fully inflated and stretched. Alternatively, maximal inspiration recruits alveoli that are normally less well ventilated, less well perfused, and hence of lower Vc. The increased cardiac output of exercise dilates capillaries and recruits capillaries, hence increasing Vc. All these processes have little effect on Dm because it is probable that NO uptake is largely independent of Vc.
Variation in DlNO between breaths.
It was not possible to perform a complete breath-by-breath analysis for Dl NOss. Although we could measure Fa NO on a breath-by-breath basis and inhaled NO would not have varied between breaths, it was not possible to calculate the mixed exhaled NO on a breath-by-breath basis. To do so would have needed exhaled flow or volume averaging of the exhaled NO signal, which is not possible with the analyzer in therapeutic mode. However, calculating Dl NOss by using Fe NO and breath-by-breath Fa NO does not show any alteration of Dl NOss with duration of inhalation or number of breaths (Fig. 3). Despite NO being vasoactive, there appears to be no evidence that breathing it alters the measurement of Dl NOss.
Critique of Method
As with Dl COss, Dl NOss suffers from the inherent inaccuracy of estimating Fa NO. We found a steep alveolar phase for exhaled NO (Fig. 2), and the place on the curve chosen as a representative value for Fa NOinevitably influenced Dl NOss to a great extent. Dl NOsb measurements are therefore more reproducible (4). Dl NOss is more practical for exercise measurements, and indeed we studied Dl NOss on exercise because at rest Dl COss is greatly affected by regional inhomogeneity in diffusing capacity, and we anticipated similar problems with Dl NOss. It would have been of interest to monitor Dl NOss continuously during increasing exercise and to have several replicates per subject. However, this would have involved exposure to 5,000 ppb NO for many minutes over several days, and we thought that this could be unsafe given that there is uncertainty regarding the safe exposure concentrations for NO. This concentration was chosen because it is stable in air but sufficiently high for interference from endogenous NO not to be a problem. The steady-state method is more applicable to use during artificial ventilation in patients and for animal work.
Comparison With Other Work
All three groups who have made Dl NOmeasurements have applied the classic Roughton and Forster model (1/Dl = 1/Dm +1/θVc) originally derived for CO transfer to NO transfer, although each group has made different assumptions, leading to differing values for Dm and Vc. Guenard et al. (8) assumed that, because the rate of reaction of NO with Hb is extremely rapid compared with the membrane conductance (Dm), Dl NO approximates to DmNO and is therefore twice DmCO because the ratio of water solubility to the square root of molecular weight for NO is twice that for CO. This value for DmCO is entered into the equation. By their reasoning, Dl NO is entirely independent of capillary blood volume. Using this approach yields lower figures for Dm and higher values for Vc than we have obtained (see Table 3). Their method and ours use upper and lower limits for θ and give correspondingly lower and upper limits for Dm and Vc. All the groups have assumed that θno is independent of alveolar Po 2, although there is no in vitro evidence for this given the technical difficulties due to the rapid reaction of NO and O2. However, our laboratory (3) has previously found some O2 variability of Dl NO over the physiological Po 2 range and presumed this to be due to alteration in Dm or Vc.
Values for DlNO-to-DlCOratio and the reaction resistance.
The ratio of Dl NO to Dl CO varies between 3.9 and 4.8, depending on which estimates of Dl are used. These estimates are rather lower than those made using single-breath estimates (4, 7), perhaps because of the exercise-induced increase in Dl CO.
Schuster and Heller (9) extended Guenard's formula and obtained the reaction resistance/total transfer resistance = Dl CO/θcoVc = 1–2/(Dl NO/Dl CO). With the use of Heller and Schuster's formula (9), which assumes that θno is infinity, the reaction resistance to CO uptake from our data of 1–2/(Dl NO/Dl CO) gives a figure of 0.48, which is close to their estimate (0.4). However, if the reaction resistance is calculated as Dl CO/(θco × Vc) using our value for Vc (Table 3) and using the O2-dependent term for θco (see calculation of Dm and Vc above), then a figure of 45.6/(158 × 4.1 × 10−3 × 100) = 0.7 is obtained. It is clearly crucial to know the in vivo value of θno to correctly calculate the reaction resistance to CO uptake.
Directions for Further Work
It would be important to confirm our observations by making measurements at rest, increasing levels of exercise and varying Po 2. A simultaneous rapidly responding flow or volume channel would allow breath-by-breath measurement and also the study of within-breath variation of Dl NO with Va. The interpretation of Dl NO and indeed Dl CO crucially depends on how values for θ obtained in the laboratory relate to the situation in the pulmonary capillary. One approach would be to measure Dl NO and Dl COsimultaneously while varying the hematocrit. If 1/Dl is plotted as a function of 1/Hb for both gases, then the ratio of slopes is θno/θco. These experiments would be best performed in an isolated lung preparation. Studies of Dl NO/Dl CO in anemia in humans have given conflicting results, no doubt because chronic anemia affects other aspects of gas transfer than hematocrit (2). In an isolated lung model, it should be possible to keep these factors constant.
Address for reprint requests and other correspondence: C. Borland, Dept. of Medicine, Hinchingbrooke Hospital, Huntingdon, Cambs. PE18 8NT, UK (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2001 the American Physiological Society
Worked Example Using Data for Subject 1
Subject 1 was 45 yr old and weighed 80 kg. The experiment was performed on a day when the laboratory temperature was 24°C and Pb was 752 Torr. Saturated vapor pressure at 37°C was assumed to be 47 Torr.
Vt = 2,200 ml, RR = 20.3, Fe CO2 = 4.8%, Fa CO2 = 5.7%, Fi CO = 0.106%, Fe CO = 0.067%, Fi NO = 5,260 ppb, Fe NO = 1,467 ppb, Fa NO = 1,200 ppb, Hb = 15.3 g/dl, COHb = 1%, metHb = 0.3%.
Dead space calculations.
Diffusing capacity calculation for DlCO.
Using predicted anatomic dead space (a DCO 1 in Bates et al.) Equation A3Using CO2 dead space [(e) in Bates et al.]
Diffusing capacity calculation for DlNO.
Using the predicted anatomic dead space for NO
Using end-tidal Fe NO from the expired NO curve (Fig. 2) as an estimate of Fa NO
Equation A4 For dead space and Dl calculations, the respiratory quotient was assumed to be 1 (see results).
Calculation of Dm and Vc.
Calculations were made as follows Equation A5 Equation A6 Equation A7For derivation of these equations, see Ref. 3. The value of θno (5) was taken as 4.5 ml · min−1 · Torr−1 · ml−1(1,500 mmol · min−1 · kPa−1 · l−1) and θco (7) calculated from 1/θco = 1.3 × 10−3 + 4.1 × 10−3 was taken as 100 Torr. For both CO and NO, correction for Hb was made by multiplying θ by Hb × (1 − COHb − metHb)/14.6.
For subject 1, we used the estimate of Dl CO using the CO2 dead space and Dl NO using end-tidal NO as an estimate of alveolar NO For calculation of DmCO, the assumption (8) is made that θno is infinity and Vc was obtained by inserting this value for DmCO into Eq. 7