## Abstract

The aim of the present study was to determine: 1) whether 40–50 ppm nitric oxide (NO) increases diffusing capacity of the lung for NO (*D*_{L,NO}) and carbon monoxide (*D*_{L,CO}), membrane diffusing capacity for CO (*D*_{m,CO}) and pulmonary capillary blood volume (*V*_{c}); 2) the actual number of tests required to provide a reasonable estimate of *D*_{L,NO}, *D*_{L,CO}, *D*_{m,CO} and *V*_{c}; and 3) repeatability of these parameters using the single-breath *D*_{L,NO}–*D*_{L,CO} method.

In total, 31 subjects performed five single-breath hold manoeuvres at rest, inhaling 43±3 ppm NO together with a standard diffusion mixture.

*D*_{L,NO} (*D*_{m,CO}) remained unchanged from the first to fifth trial. However, compared with the first trial, *D*_{L,CO} and *V*_{c} had decreased by the fourth (−4±5%; 95% confidence interval (CI) = −5– −2%) and third trial (−5±7%; 95% CI = −7– −2%), respectively. Repeatability over five trials was 17, 3 and 7 mL·min^{−1}·mmHg^{−1} for *D*_{L,NO}, *D*_{L,CO} and *D*_{m,CO}, respectively, and 13 mL for *V*_{c} when *D*_{m,CO} = *D*_{L,NO}/2.42.

In conclusion, nitric oxide inhaled during sequential single-breath manoeuvres has no effect on diffusing capacity of the lung for nitric oxide and, thus, membrane diffusing capacity for carbon monoxide. Since more than two and three trials will lower pulmonary capillary blood volume and diffusing capacity of the lung for carbon monoxide, respectively, the average value of only two properly performed trials is suggested.

The equation of the diffusing capacity of the lung for carbon monoxide (*D*_{L,CO}) has been classically described as: where *D*_{m,CO} represents pulmonary membrane diffusing capacity for carbon monoxide (CO) and ΘCO is the specific blood transfer conductance for CO. *V*_{c} represents pulmonary capillary blood volume 1. Membrane resistance (1/*D*_{m,CO}) and red cell resistance (1/ΘCO·*V*_{c}) usually contribute equally to the overall diffusive resistance across the lung 2, although this has been debated 3. To obtain *D*_{m,CO} and *V*_{c}, *D*_{L,CO} has been traditionally measured at two different levels of alveolar oxygen tension (*P*_{A,O2}), ∼13.3–16.0 kPa (∼100–120 mmHg) and ∼79.8 kPa (∼600 mmHg). For each level of *P*_{A,O2}, 1/*D*_{L,CO} is then plotted on the y-axis and 1/ΘCO is plotted on the x-axis. A line is then drawn through two points and the x-intercept (1/*D*_{m,CO}) and slope (1/*V*_{c}) can be solved. This two-step method can be time consuming and uncomfortable to perform, especially during exercise.

However, over the past 15 yrs, the measurement of diffusing capacity of the lung using the transfer gases nitric oxide (NO) and CO together permits one to obtain *D*_{m,CO} and *V*_{c} in a single-breath manoeuvre, thus allowing a similar distribution of the two gases and reducing the number of measurements and testing time by half 4, 5. The velocity constant of the combination of NO with haemoglobin is about 280 times faster than that of CO 6, and thus the specific blood transfer conductance for NO (ΘNO) is so large that the red cell resistance to NO (1/ΘNO) approaches zero 7. Therefore, diffusing capacity of the lung for nitric oxide (*D*_{L,NO}) equals the membrane diffusing capacity for NO (*D*_{m,NO}), and is independent of *V*_{c} and haemoglobin concentration 8. Others have made the same assumption that (1/ΘNO) is negligible 9–14, and it was recently determined that a nonzero 1/ΘNO would not be able to explain their experimental data 12. Therefore, these data suggest that *D*_{L,NO} is a good measure of *D*_{m,NO}. Given that the molecular weight of CO and NO are 28 and 30 g·M^{−1}, respectively, and solubilities (Bunsen coefficients) of CO and NO in plasma at 37°C are 0.0215 and 0.0439 15, respectively, the diffusivity of NO, which is the solubility of NO divided by the square root of the molecular weight (MW) of NO (), is ∼1.97 times greater than that of CO and, thus, the theoretical factor between membrane diffusing capacities for NO and CO is: Indeed, the solubility of either gas will depend upon the composition of the fluid that the gas has to diffuse through. If the fluid changes composition, the relative solubilities may well be different pre- and post-exercise. Nevertheless, data obtained from sick and healthy patients performing rebreathing manoeuvres suggest that the actual ratio of *D*_{L,NO} to *D*_{m,CO} is ∼2.42 11, 12. The larger ratio may be due to a higher than assumed NO solubility in plasma as well as NO facilitated diffusion 12.

Several researchers have previously obtained *D*_{L,NO} from single-breath 4, 5, 9, 10, 13, 14, 16–19 or rebreathing manoeuvres 11, 12, along with the simultaneous measurement of *D*_{L,CO} to obtain *D*_{m,CO} and *V*_{c}. Since brief exposure to NO does not interfere with physiological function 11, 12, 20, 21, it seems pertinent to use NO as a test gas to assess lung diffusion capacity. However, there is still debate as to whether inhalations of high NO concentrations (∼40–50 ppm) during sequential single-breath manoeuvres can affect the pulmonary microcirculation. At those NO concentrations, pulmonary vasodilation may occur, increasing *D*_{L,CO}, *V*_{c} and perhaps *D*_{L,NO} (*D*_{m,CO}) 22. Therefore, the first objective of the present study was to determine whether five sequential single breath-hold manoeuvres inhaling ∼40 ppm of NO increase *D*_{L,CO}, *D*_{L,NO} (and thus *D*_{m,CO}) and *V*_{c}. The current hypothesis was that five repeated inhalations of 40–50 ppm of NO would not increase *D*_{L,CO}, *D*_{L,NO} (*D*_{m,CO}) or *V*_{c}.

Furthermore, the recent 2005 American Thoracic Society (ATS)/European Respiratory Society (ERS) task force guidelines of standardisation of the single-breath determination of CO uptake in the lung 23 mentioned that more research is needed to determine the actual number of tests required to provide a reasonable estimate of the average *D*_{L,CO}. Therefore, the second objective of the current study was to determine the actual number of tests required to provide a reasonable estimate of not only *D*_{L,CO}, but *D*_{L,NO} (*D*_{m,CO}) and *V*_{c} for a given person. The present authors' hypothesis was that an average of three measurements would be needed to obtain a reasonable estimate of all those parameters.

The third objective was to determine the repeatability of lung diffusing capacity and its components over five trials in a given patient using the newer single-breath method of CO and NO. This would help to decide whether a change in an observation of *D*_{L,NO}, *D*_{L,CO}, *D*_{m,CO} and *V*_{c} represents a real clinical change in the pulmonary system or just measurement error. The current authors' hypothesis was that the repeatability for *D*_{L,NO}, *D*_{L,CO} and *D*_{m,CO} would be 10, 2 and 6 mL·min^{−1}·mmHg^{−1}, respectively, and 10 mL for *V*_{c}.

## METHODS

### Subjects

In total, 31 healthy subjects were recruited (15 females, 16 males) and all completed the study (aged 33±9 yrs; weight 68.6±12.5 kg; height 169.9±9 cm). All subjects were nonsmokers. These subjects had normal resting spirometry function (forced expiratory volume in one second (FEV_{1)} >80% predicted, and FEV_{1}/forced vital capacity (FVC) >0.70) and no history of cardiopulmonary disease. Each subject was required to come into the laboratory on one occasion only.

### Single-breath apparatus and technique

Volume and gas calibration of the Ergocard and the Hyp'Air lung diffusion system (Medisoft, Dinant, Belgium) were performed prior to each testing session. The subjects breathed through a three-way pneumatic valve developed by Medisoft. A dead space washout volume of 900 mL was allowed, and an expired sample volume of 900 mL was collected. The instrument dead space was measured at 140 mL (including the mouthpiece, valve and filter dead spaces). Anatomical dead space (mL) was estimated as bodyweight in kg×2.2 23. The concentrations of inspiratory gases were 0.295% CO, 9.96% He, 20.98% O_{2} and balance N_{2} for gas mixture one, and 1,000 ppm NO and balance N_{2} for gas mixture two. A third mixture of 77 ppm NO and balance N_{2} was used for calibration purposes only. For the single-breath manoeuvre, an inspiratory bag was filled with 5–8 L depending on the subject's FVC using the first two gas mixtures. Once the mixtures were injected into the inspiratory bag, the concentration of CO, He, NO, and O_{2} were analysed simultaneously over 30 s by gas analysers. The injection of the various gas mixtures into the inspiratory bag resulted in approximate inspired concentrations of CO, He, NO and O_{2} at 0.30%, 9%, 40 ppm and 20%, respectively. The types of gas analysers used for measuring inspired and expired gas mixtures have been reported elsewhere 13. Inspired volume was measured, corrected for instrument and anatomical dead space, and converted to standard temperature, pressure and dry conditions. Breath-holding time was calculated using the method of Jones and Meade 24.

### Calculation of diffusion capacities, *D*_{m} and *V*_{c}

Diffusion capacities for NO and CO were calculated simultaneously from the exponential disappearance rate of each gas with respect to He using the method by Jones and Meade 24. The formulae for calculating *D*_{L,CO} can be found in the 2005 ATS/ERS guidelines 23. All results were standardised to a haemoglobin (Hb) concentration of 14.6 g·dL^{−1} for males and 12.0 g·dL^{−1} for females, and a *P*_{A,O2} of 13.3 kPa (100 mmHg) by inserting these values into the following formula by Roughton and Forster 1: where 1/ΘCO was 1.426 and ΘCO was 0.701 mL·min^{−1}·mmHg^{−1} for males, 1/ΘCO was 1.594 and ΘCO was 0.627 mL·min^{−1}·mmHg^{−1} for females and *P*_{O2} was partial pressure of oxygen. A *D*_{L,NO} to *D*_{m,CO} ratio of 2.42 11, 12, 14 and 1.97 5, 9, 10, 13, 19 was used as the theoretical ratio of *D*_{L,NO} to *D*_{m,CO} during single-breath manoeuvres since those ratios have both been used in the past. The ratio of 2.42 has been determined recently during rebreathing manoeuvres at rest and during exercise 11, 12, which can result in *D*_{m,CO} values that are more in line with the current normative values 25. Therefore, due to the two different *D*_{L,NO} to *D*_{m,CO} ratios reported in the literature, the current authors reported two different *D*_{m,CO} values and two different *V*_{c} values.

A breath-holding time of 4–5 s was chosen because it has been shown that *D*_{L,CO} values were not different with a 3- or 5-s breath hold compared with a 7- and 10-s breath hold 26 and a breath-holding time of ∼9 s would result in a less than detectable amount of expired NO 4. The present authors did not account for NO back pressure in the calculations since exhaled NO concentrations at rest are negligible, ranging 11–66 ppb (0.011–0.066 ppm) 27, which tend to decrease during exercise 28. The amount of NO back pressure then would minimally affect *D*_{L,NO} calculations as the measurements were carried out in the ppm range, which is ∼75–500 times larger than the exhaled NO concentration at rest or during exercise after a single-breath inspiration of 40–60 ppm NO. Accounting for CO back pressure is also negligible as it has been shown that 2 min between tests virtually eliminates all the CO gas from the lungs in healthy subjects 29. As the recent ATS/ERS guidelines recommend a minimum of 4 min between *D*_{L,CO} measurements 23, subjects performed the single-breath manoeuvre with a minimum of 4.5 min between tests. Five sequential diffusion capacity tests were performed.

### Statistical analyses

A one-way repeated measures ANOVA was used to determine if there were significant differences in *D*_{L,CO}, *D*_{L,NO}, *D*_{m,CO} and *V*_{c} between the five trials. A Tukeys/Kramer *post hoc* comparison test was used to see where the differences in the five trials occurred. In addition, the variables height, age and weight were examined by forward stepwise multiple regression to determine which and what combination most predicted *D*_{L,NO} values at rest. The measurement error for *D*_{L,NO}, *D*_{m,CO}, *D*_{L,CO} and *V*_{c} was calculated as the square root of the residual mean square error (which is also called the within-subject standard deviation) obtained from the one-way repeated measures ANOVA 30. The repeatability of each variable was then obtained multiplying the within-subject standard deviation by 2.77 30. A p-value of <0.05 was considered statistically significant.

## RESULTS

Of the 31 subjects, 29 had normal spirometry function (table 1⇓) based on the fact that FEV_{1} was >80% pred for all subjects, and the FEV_{1}/FVC was >0.7 for all but two subjects (FEV_{1}/FVC 0.67 and 0.68, respectively). The single breath-hold manoeuvres, on the whole, were performed adequately (table 2⇓). Breath-hold time was consistent within 0.3 s for all five trials, and the average inspired volume was always >90% of the FVC. The average alveolar volume was also maintained within 0.2 L for all trials.

The data show that *D*_{L,NO} and, thus, *D*_{m,CO} remained unaltered from trial one to trial five (table 3⇓). However, compared with the first trial, *D*_{L,CO} and *V*_{c} significantly decreased by the fourth (−4±5%; 95% confidence interval (CI) −5– −2%; p<0.05) and third trial (−5±7%; 95% CI −7– −2%; p<0.05), respectively. When *D*_{m,CO} = *D*_{L,NO}/2.42, the per cent predicted based on age, height and sex was 111±31% (range 62–164%; p<0.05), but when *D*_{m,CO} = *D*_{L,NO}/1.97, the per cent predicted was 136±39% (72–202%; p<0.05). For *V*_{c}, when *D*_{m,CO} = *D*_{L,NO}/2.42 the per cent predicted based on height and sex was 116±19% (76–155%; p<0.05), but when *D*_{m,CO} = *D*_{L,NO}/1.97 the per cent predicted for *V*_{c} was 100±16% (66–131%; p>0.05). Therefore, when *D*_{m,CO} = *D*_{L,NO}/2.42, the *D*_{m,CO} is reduced closer to normative values and the *V*_{c} is increased above normative values. The opposite occurs when *D*_{m,CO} = *D*_{L,NO}/1.97 as the *D*_{m,CO} is largely increased above the predicted value, but *V*_{c} is reduced to 100% pred.

The measurement error and repeatability of five trials are presented in table 4⇓. The repeatability represents the critical value in which a clinically measurable change in a given patient occurs between sessions. From the repeated measures ANOVA, a real clinical change in a patient's pulmonary diffusion capacity and its components was 17.2 and 3.2 mL·min^{−1}·mmHg^{−1} for *D*_{L,NO} and *D*_{L,CO}, respectively. When *D*_{m,CO} = *D*_{L,NO}/2.42, a real clinical change for *D*_{m,CO} and *V*_{c} were 7.1 mm·min^{−1}·Hg^{−1} and 13 mL, respectively. When *D*_{m,CO} = *D*_{L,NO}/1.97, a real clinical change for *D*_{m,CO} and *V*_{c} was any change >8.7 mL·min^{−1}·mmHg^{−1} and 9.8 mL, respectively.

The variables age, weight and height were examined by forward stepwise multiple regression to predict *D*_{L,NO} values at rest. The only variable that appreciably affected *D*_{L,NO} was height. Predicted *D*_{L,NO} in mL·min^{−1}·mmHg^{−1} = 2.0164×height in cm–175.63 (r^{2} = 0.39; see = 23.3). Adding the other two variables to the equation did not increase the coefficient of determination significantly. Therefore, *D*_{L,NO} at rest was best predicted by height.

## DISCUSSION

The present study showed that repeated inhalations of 40–50 ppm NO during single breath-hold manoeuvres does not increase *D*_{L,NO}, *D*_{m,CO}, *D*_{L,CO} or *V*_{c}. In fact, five sequential breath-hold manoeuvres did not change *D*_{L,NO} and, thus, *D*_{m,CO}. However, *D*_{L,CO} significantly decreased by the fourth trial, and *V*_{c} significantly decreased by the third trial. The drop in *D*_{L,CO} by the fourth trial due to progressive increase in carboxyhaemoglobin was similar to that reported elsewhere 33. As such, the data show that the actual number of tests required to provide a reasonable estimate of *D*_{L,NO}, *D*_{L,CO}, *D*_{m,CO} and *V*_{c} during properly performed manoeuvres is two trials. The average of the two trials should then be reported. Any more than two properly performed manoeuvres will lower *V*_{c}, and more than three will lower *D*_{L,CO}. Therefore, two properly performed manoeuvres are sufficient to obtain all the components of lung diffusion capacity using the *D*_{L,NO}–*D*_{L,CO} method.

There has been some concern that NO may affect the pulmonary microcirculation due to its vasoactive properties 22. One study showed that there is a small nonsignificant increase in *D*_{L,CO} in the presence of NO 17. However, the current authors have shown that brief exposure to NO does not interfere with physiological function. Other human 11, 12, 20, 21 and animal studies 34 are in agreement with the present study. Rebreathing 20–40 ppm NO over 16 s to 10 min resulted in no significant changes to oxygen uptake, arterial oxygen tension, alveolar–arterial oxygen tension difference, *D*_{L,CO}, *D*_{m} or *V*_{c} in sick and healthy subjects at rest or during exercise 11, 12, 21. In mechanically ventilated rabbits, *D*_{L,NO} values remained unchanged despite inspiratory NO concentrations varying from 10–800 ppm 34. Furthermore, the pulmonary toxicity of inhaled NO at concentrations of ∼40 ppm over a prolonged period of time is minimal 35.

There is concern that different *P*_{A,O2} may slightly alter *D*_{L,NO} 36. However, more recent data has shown that varying the *P*_{A,O2} in those with or without pulmonary disease does not affect *D*_{L,NO} 11, 12, implying that the combined *D*_{L,NO}–*D*_{L,CO} method can be used in hypoxaemic patients. Therefore, taken together, there is no reason to refrain from using NO and CO concurrently as a test gas to assess lung diffusion capacity.

The clinical implication of using both NO and CO concurrently in research and medical practice is that scientists and clinicians can immediately partition and quantify the components of lung diffusing capacity in a subject from a single 4-s breath-hold manoeuvre that requires minimal effort on the part of the patient, while simultaneously being able to pinpoint which component (*V*_{c}, *D*_{m}) is causing low (or high) total lung diffusion capacities. The ability to estimate *D*_{m,CO} and *V*_{c} from one-step simultaneous measurement of *D*_{L,NO} and *D*_{L,CO} represents significant conceptual advantages. One conceptual advantage of the *D*_{L,NO} technique is that with the standard Roughton–Forster method, cardiac output can vary between measurements of *D*_{L,CO} at different O_{2} tensions, which then have to be interpolated to obtain *D*_{L,CO} at two O_{2} tensions, but at the same cardiac output 12. With the *D*_{L,NO} method, all measurements are obtained at the same cardiac output and O_{2} tension; no interpolation is necessary. Another conceptual advantage is that with the traditional method, the distribution of CO gas in the lungs may be different at two different O_{2} tensions, but with the *D*_{L,NO} method only one inspiration is required, which results in a similar distribution of NO and CO gases. A third conceptual advantage is that with the traditional method, there is systematic underestimation of *V*_{c} and an overestimation of *D*_{m} since the inspiration at two different O_{2} tensions affects alveolar–capillary membrane diffusion 37. The *D*_{L,NO}–*D*_{L,CO} method avoids this error altogether and should improve the accuracy of estimated *D*_{m,CO} and *V*_{c}.

The ratio *D*_{m,CO} = *D*_{L,NO}/2.42 gives a better estimate of *D*m. When the present measured values were compared against the published norms, using the traditional two-step Roughton–Forster method 25, the predicted *D*_{m,CO} was 136% pred when *D*_{m,CO} = *D*_{L,NO}/1.97, but only 111% pred when *D*_{m,CO} = *D*_{L,NO}/2.42. However, the better per cent predicted values for *V*_{c} occurred when *D*_{m,CO} = *D*_{L,NO}/1.97. Nonetheless, since the traditional Roughton–Forster method underestimates *V*_{c}, the current authors feel that the predicted *V*c values by Zanen *et al.* 25 slightly underestimate *V*_{c}. Therefore, the best compromise is to estimate *D*_{m} and *V*_{c} from a single breath using the *D*_{L,NO}–*D*_{L,CO} method and the formula *D*_{m,CO} = *D*_{L,NO}/2.42.

It was also important to clarify the repeatability of lung diffusion capacity and its components over five trials from the single-breath *D*_{L,NO}–*D*_{L,CO} method. The repeatability allows clinicians to identify a true clinically meaningful change from measurement error. The difference between any two measurements for the same subject is expected to be <2.77 multiplied by the within-subject standard deviation; therefore, the repeatability was defined as 2.77 multiplied by the within-subject standard deviation obtained from the ANOVA 30. Table 4⇑ shows a true measurable clinical change in *D*_{L,NO} and *D*_{L,CO} as an absolute change of >17 and 3 mL·min^{−1}·mmHg^{−1}, respectively. In addition, a true clinical change in *D*_{m,CO} and *V*_{c} is an absolute change of >7 mL·min^{−1}·mmHg^{−1} and 13 mL, respectively. Any change that is less than these values is considered a true measurement error. Indeed, *D*_{m,CO} equals *D*_{L,NO} divided by a fixed value, and *V*_{c} is derived from the *D*_{L,NO} and *D*_{L,CO}; therefore the repeatability of *D*_{m} and *V*_{c} can be calculated from (or explained by) the repeatability of the *D*_{L,NO}.

Based on ATS and ERS criteria, the average value of two trials whose difference in diffusing capacity is within 10% of each other is considered acceptable. The present authors looked at the average *D*_{L,NO} and *D*_{L,CO} for all 31 subjects over five trials, and the repeatability data of this study are in agreement with the ATS/ERS guidelines as *D*_{L,NO} (and thus *D*_{m,CO}) and *D*_{L,CO} are found to be 10%. However, since the variability of the parameters *D*_{L,NO} and *D*_{L,CO} were independent of the magnitude of the measurement, the results invalidate the use of percentage value to describe the repeatability. Using a percentage will lead to underestimation of variability in low values and overestimation for high values. Others studies have also suggested using an absolute value rather than a percentage, since the diffusing capacity was also independent of the magnitude of the measurement 38, 39. Therefore, the current authors recommend using an absolute difference rather than a percentage as alternative criteria for repeatability. It is also recommended to report the average of two trials when the absolute difference between the two measurements is within 17 mm·min^{−1}·mmHg^{−1} for *D*_{L,NO}, 3 mm·min^{−1}·mmHg^{−1} for *D*_{L,CO}, and 13 mL for *V*_{c}.

It is believed that the present paper is the first to actually quantify important measurable clinical changes in the parameters obtained from the single-breath *D*_{L,NO}–*D*_{L,CO} method.

### Conclusion

Small amounts of nitric oxide inhaled during sequential single-breath manoeuvres have no effect on lung diffusing capacity for nitric oxide and, thus, membrane diffusing capacity for carbon monoxide. The recommended ratio is membrane diffusing capacity for carbon monoxide = lung diffusing capacity for nitric oxide/2.42. As more than two and three single-breath manoeuvres will lower pulmonary capillary blood volume and lung diffusing capacity for carbon monoxide, respectively, the average value of the first two trials are recommended to provide a reasonable estimate of lung diffusing capacity for nitric oxide, lung diffusing capacity for carbon monoxide, membrane diffusing capacity for carbon monoxide and pulmonary capillary blood volume.

- Received December 12, 2005.
- Accepted February 24, 2006.

- © ERS Journals Ltd