INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Measurement of exhaled nitric oxide (FE_NO) is the first noninvasive and practical method for assessing airway inflammation (1). Moreover, exhaled nitric oxide measurements appears to be a more sensitive indicator of severity of asthmatic inflammation than serum eosinophil cationic protein (ECP) and soluble interleukin 2 (IL-2) (10), and is useful in monitoring the response to antiinflammatory medications (11). Thus exhaled nitric oxide may be used as a tool to monitor inflammation in asthma and to titrate antiinflammatory therapy.

However, reported FE_NO values for healthy children and children with asthma have shown wide variation depending on the measurement technique used by different investigators, making comparison between different laboratories impossible. Currently identified as factors most affecting the measurement of FE_NO levels are expiratory flow rate, portion of flow profile (peak versus plateau values taken as FE_NO level), and contamination by significantly higher levels of NO produced in the nasal cavity and paranasal sinuses (14, 15). These differences in measuring techniques underscore the need for standardized procedures (16, 17).

While the single-breath constant flow technique is widely used by investigators studying adults, there is less consensus for its use in pediatric studies. Children have varying lung sizes and differ in their ability to cooperate and maintain the steady exhalation flow used in this method. Currently, there is limited experience in using the single-breath technique to measure FE_NO values in children (10, 18). This method involves measuring FE_NO during a constant expiratory flow and eliminates NO contamination from nasopharyngeal sources by exhaling against a constant pressure, which closes the velum. Measurement of FE_NO at low flow rates greatly amplifies the NO signal and provides better discrimination between healthy subjects and those with disease states (15), and avoids measurement near the detection limits of current NO analyzers (21). However, low flow rates require exhalation for a longer period of time in order to achieve plateau FE_NO values. There is also a need to study different collection techniques (i.e., offline techniques where exhalate is collected for delayed analysis) in order to increase the potential usefulness of FE_NO measurement as a tool for diagnosis and monitoring of therapeutic response in a variety of clinical settings (i.e., hospital wards, home, or school).

The aims of our present investigation were to determine the most appropriate flow rate for the single-breath technique in children and to obtain normal values that can be used for controls in further studies. We measured FE_NO values using a single-breath technique at eight flows in 32 healthy adolescents (15 to 18 yr of age), collecting both online and offline specimens (see METHODS). We aimed to determine which flows are easy to sustain, generate reproducible values for FE_NO, and provide good correlation between offline and online measurements.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

The 32 subjects were 15- to 18-yr-old healthy adolescents. Inclusion criteria were as follows: no evidence of lower respiratory tract infection (history of cold, cough, and fever for 10 d before study, no evidence of nasal obstruction, tenderness in the area of the paranasal sinuses, temperature > 100° F, wheezing, or crepitations on chest auscultation); no history of any chronic respiratory or cardiovascular illness including asthma, cystic fibrosis, or immunodeficiency syndromes; and no history of seasonal or perennial allergies. Candidates also had no history of drug intake during the 72 h preceding the study, no reported history of smoking within the past year, and no evidence of airway obstruction on spirometric testing (FEV₁ > 80% predicted according to Crapo and coworkers) (22). The protocol was approved for human studies by the Institutional Review Boards of the University of Florida, Nemours Children's Clinic, and Wolfson Children's Hospital, as well as their research and ethics committees. A written informed consent was obtained in all cases.

Measurement of Lung Function

Patients performed spirometry (forced vital capacity [FVC] and forced expiratory volume in 1 s [FEV₁]) with a Fleish (Lausanne, Switzerland) pneumotach spirometer with an integral temperature sensor (Multi-spiro-PC; Medical Equipment Design, Laguna Hills, CA) connected to an IBM Pentium computer with an integrator to convert flow rate into volume. The best of three trials, as determined by the highest FEV₁, were used to determine the baseline FEV₁.

FE_NO Measurement Procedure

FE_NO was measured with a chemiluminescent nitric oxide analyzer (NOA 280; Sievers Instruments, Boulder, CO). The sensitivity of the analyzer for measurement of gas-phase NO is less than 1 ppb by volume. The NOA device was calibrated at 0 and 50 ppm in accordance with the manufacturer recommendations. As subjects inhaled ambient air, ambient NO concentrations were measured at the time of each test. If the ambient NO concentration was high (> 30 ppb), the data were discarded (16). After the first 10 subjects, and the finding that ambient NO levels vary considerably from day to day, we used an NO scrubber (Window-Cator cannister type N-P100; Mine Safety Appliance, Pittsburgh, PA) to filter ambient air before inhalation, which resulted in inspired air containing < 1 ppb of NO.

The FE_NO measurement was obtained with the subjects in a standing position. After inhaling to total lung capacity the subjects exhaled through a mouthpiece attached to a one-way valve containing two sampling ports. Nitric oxide was sampled directly in the analyzer (at a flow rate of 250 ml/min) through a Teflon side arm tube attached to one of the sampling ports. The tubing is 60 in long with an internal diameter of 1/16 of in. Exhalation mouth pressure was measured by a pressure transducer in the analyzer via the second sampling port. Both pressure and NO were displayed simultaneously on the front panel of the analyzer and on a computer attached to the RS-232 output of the NOA. Data were stored and analyzed on the computer, using NO analysis software (Sievers Instruments).

The flow rates were achieved by the placement of expiratory resistances in the exhalation circuit and by asking the subject to exhale at a constant mouth pressure, which was displayed and readily visualized by the subject on a computer screen. The pressure bar remained red until the target pressure was obtained, at which time it changed to green. If the flow dropped below or increased above the desired range, the green light would again change to red. Of more importance, the plateau nitric oxide level would not be maintained; therefore, the trial would be considered inadequate. However, this was the exception rather than the rule. The set point for color change from green to red is within 2% of the desired flow rate. With the color change of the pressure bar and visual examination of the plateau, we feel confident that flows were maintained in the desired range. Two to four trials were required for subjects to understand how to perform the maneuver before formal testing.

Nitric oxide measurements were made for each child using two different mouth pressures (10 and 20 cm H₂O) and four different resistors (16-, 18-, 20-, and 22-gauge needles giving eight separate flows). Needles were calibrated with a mass flow meter in series with a Baratron pressure gauge and a Sievers breath kit (modified Hans Rudolf valve from Sievers Instruments). The combinations of needles and pressures yielded corresponding expiratory flow rates of 46, 31, 23, 15, 10, 7, 5, and 4 ml/s. These expiratory flow rates were the sum of the analyzer flow and that measured by an expiratory pneumotachygraph. These flow rates were chosen since they were in the range acceptable to most children. In addition, our previous experience suggested that these flow rates may be optimal, especially in younger children. After inhalation to total lung capacity, the children immediately exhaled into the mouthpiece. The mouthpiece has a 0.2-µm pore size bacterial filter to prevent contamination within the valve. The total deadspace of the mouthpiece is 4 ml. The exhalation proceeded until a stable plateau was reached. Three FE_NO plateau measurements that varied by < 10% were taken for each different flow rate and the average noted. Subjects were also asked to comment on the ease of performing each of the maneuvers.

Flow Acceptability Range

In our initial 10 subjects, FE_NO values were obtained at eight different flow rates (46, 31, 23, 15, 10, 7, 5, and 4 ml/s). In these children, velum closure and thus separation of nasal from lower respiratory NO was confirmed by placement of a CO₂ probe approximately 3 cm in the left nostril. Velum closure was confirmed by no rise in CO₂ value on exhalation. The data obtained from the first 10 subjects were analyzed to determine ease of maintaining expiratory flows, time to plateau, and correlation between online and bag samples (samples of exhaled air collected in Tedlar bags for later analysis) of FE_NO values (19, 20).

Flow Dependency

On the basis of our findings in the first 10 subjects, the next 22 subjects were studied at the four higher flow rates and without the use of nasal CO₂ probes. We elected to use these four flows because of the ease with which subjects were able to generate a plateau pressure in a reasonable time. Furthermore, at these higher flows FE_NO values demonstrate much less variability. Moreover, we felt that these flows are likely to yield values, which allow for better discrimination between asthmatic and healthy children.

Offline (Bag) Method

Exhaled gas was collected in Tedlar bags, simultaneously with online measurement, by placing the collection bag as a reservoir at the exhalation port (Figure 1). Bags were sealed and analysis of FE_NO from the bags was completed within 2 h by aspirating the bag, using the analyzer sample line. The contents of the bag were drawn into the NO analyzer (250 ml/min) through Teflon tubing (5-8 cm long, i.d. 5.5 mm, o.d. 8 mm).

View larger version (15K): [in this window] [in a new window]   Figure 1.   Experimental setup for online and offline (bag collection) NO analysis. The computer screen provides an incentive feedback to maintain a specified pressure.

Statistical Analysis

Silkoff and coworkers reported a high correlation (r > 0.90) between pulmonary FE_NO values and flow rates (15). In this study, we also anticipated a large effect size (r > 0.50) for the correlation between flow and FE_NO and for the correlation between online and bag collection methods. On the basis of a large effect size and an  of 0.05, a sample size of 30 was required for a power greater than 80%.

The Shapiro-Wilk test was used to determine whether distributions for NO plateau values at the low flow rates tested conformed to a normal Gaussian distribution. Results indicated a departure from normality. Natural logarithmic transformations were applied to reduce skewness and kurtosis and deviation from normal distribution. The relationship between NO plateau and flow was analyzed by least-squares regression. Pearson product-moment correlation coefficients and paired Student t tests were used to examine the relationships between pulmonary FE_NO values collected by bag and those analyzed online, and between flow rate and FE_NO values. Bland-Altman analysis (23) was used to assess agreement between methods.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Thirty-two healthy 15- to 18-yr-olds (21 boys, 11 girls) were recruited. Lung function and other baseline descriptions of the group are presented in Table 1. All children were within the normal range for weight and height. Lung function tests were all within normal limits. There were no significant differences in body surface area or lung functions between males and females. Table 2 provides the mean and standard deviation for online and bag collection techniques at eight different flow rates, and Table 3 summarizes the raw data based on age and sex.

Flow Acceptability Range

Our initial 10 subjects were tested at all eight flow rates. At the four lowest flow rates, there were significant differences and poor correlation between bag and online values (r = 0.33 to 0.74; p > 0.05). In addition, subjects found it difficult to sustain these flows long enough to maintain a plateau FE_NO level. Time to plateau ranged from 6.32 to 6.62 s for the four highest flows (15-46 ml/s) and from 11.98 to 12.14 s for flows of 4 to 10 ml/s. Subsequently, flow rates between 15 to 46 ml/s were used for the next 22 subjects. The following results relate, therefore, to these four flows in all 32 subjects.

Flow Dependency

NO output (FE_NO × flow rate) was calculated by the formula presented by Silkoff and coworkers (15). Figure 2 depicts the NO output means at each flow for both bag and online collections. As found by Silkoff and coworkers, NO excretion increases as flow decreases. There were no significant differences in NO excretion between age groups at any flow or collection technique (p = 0.26 to 0.91). As with FE_NO, there was a significant difference between males and females at the four highest online flows (p = 0.015 to 0.025), but not for the lowest online flows (p = 0.54 to 0.83) or any bag flows (p = 0.11 to 0.85).

View larger version (15K): [in this window] [in a new window]   Figure 2.   Mean nitric oxide output (log transformed) versus flow for online and bag collection technique at all eight flows. Flow dependency is demonstrated. As flow decreases, FENO increases. In addition, at higher flows, bag and online values are almost identical. Online (solid squares); bag (solid triangles).

Bag versus Online Comparisons

FE_NO values increased as flow rates decreased (Figure 3) with strong correlation between FE_NO and flow rates at the four highest flows (0.85 to 0.93, p < 0.001). Using paired t tests for these same flows, no significant differences were found between bag and online FE_NO values (p < 0.09 to p < 0.83). Using Bland- Altman analysis (23), there was good agreement between online and bag measurements, as depicted in Figure 4. Intraclass correlation coefficients for repeated measurements ranged from 0.78 to 0.96 (online) and from 0.67 to 0.96 (bag) for the four lowest flows. Online intraclass correlation coefficients for the four highest flows were 0.99 for all online and bag values, except for the bag values at flow 15 ml/s, which was 0.95. 

View larger version (15K): [in this window] [in a new window]   Figure 3.   Log transformed online FENO values versus different flow rates for all ages. In general, FENO increases with age, except in 18-yr-olds. FENO values appear to peak at 17 yr. 15-yr-olds (solid circles); 16-yr-olds (solid triangles); 17-yr-olds (solid squares); 18-yr-olds (solid inverted triangles).

View larger version (19K): [in this window] [in a new window]   Figure 4.   Bland-Altman analysis plotting bias versus average of online and bag measurements at the four highest flows. As the flow rate increases, the dispersion of FENO values decreases: 46 ml/s (solid circles); 31 ml/s (solid squares); 23 ml/s (open wedges); 15 ml/s (open triangles).

Relationship of FE_NO to Other Variables

Multiple regression was used to examine the multivariate nature of FE_NO. Pulmonary function parameters, sex, flow, ambient NO, time to plateau, and age were selected as predictor variables. The data for online and offline measurements were similar and could be explained by the same model. The dependent variable was the log-transformed mean NO of the three samples. The most parsimonious model with the highest explained variance was a forward regression with regression through the origin and a probability of F-to-remove  = 0.10, R² = 0.98. The final model included the FEF_25-75, age, flow, and body surface as significant predictors of FE_NO. The regression equation was:

Age Trends

Online FE_NO values increased from 15 to 17 yr at all flow rates, but, suprisingly, decreased at age 18 yr (Figure 5). This pattern of decreased flow was consistent across all flow rates. However, using one-way ANOVA, the differences between age groups were not significant at any flow rate for either online or bag collection techniques (p = 0.26 to 0.91). There were significant differences in online FE_NO values between boys and girls (Figure 6) at the four highest flow rates (p = 0.015 to 0.025), with FE_NO values in males being higher. These differences were not significant in bag collected specimens (p = 0.11 to 0.16). No significant differences were noted between boys and girls for any of the four lowest flows, either online or bag (p = 0.48 to 0.91) (Figure 6).

View larger version (11K): [in this window] [in a new window]   Figure 5.   Log-transformed FENO values as a function of age. Values are given at a flow rate of 46 ml/s. This figure demonstrates clearly the age variable in FENO with peak levels at 17 yr.

View larger version (15K): [in this window] [in a new window]   Figure 6.   Differences between online and bag log-transformed FENO values for boys and girls in relationship to flow rates. FENO values in males were significantly higher than in females at the four higher flows. Males/online (solid squares); males/bag (solid triangles); females/ online (solid inverted triangles); females/bag (solid circles). Asterisks indicate significant differences between males and females for online values.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study represents an attempt to determine an expiratory flow rate that would be appropriate for use in children as well as an attempt to establish normal reference values for the 15- to 18-yr-olds using the constant flow, single-breath technique for FE_NO analysis. As a secondary objective, we determined whether there are age and sex differences in this reference population.

While the understanding of the physiologic basis of NO production in the airway is incomplete, there is considerable interest in NO measurement because of its relationship to airway inflammation and asthma. Studies have consistently shown that irrespective of the method used, FE_NO levels are increased in asthma compared with healthy subjects and antiinflammatory drugs such as corticosteroids reduce FE_NO concentrations, suggesting that FE_NO could be a useful index of asthma activity and treatment efficacy (24).

It is well recognized that expiratory flow rates are a major determinant of the value of the nitric oxide plateau obtained (15, 18). Silkoff and colleagues demonstrated an almost 35-fold increase in mean NO plateau values as flows decreased from 1,550 to 4.2 ml/s. (15) Franklin and colleagues, using a narrower flow range (driving pressures of 10, 15, and 20 cm H₂O), which resulted in flows of 50, 75, and 100 ml/s, found a similar relationship although less dramatic (18). We demonstrated a similar trend in our study as outlined in Figure 2. In addition, Figure 2 also demonstrates that at higher flows, the bias or mean difference between inline and bag methods was small, while at lower flows this was not the case. We also demonstrated that at extremely low flows the values obtained were inconsistent, resulting in larger relative standard deviations for both inline and bag methods. In our limited experience with children with moderate obstructive disease, these flow rates seem to be appropriate; however, we recognize that children with severe disease may have difficulty in achieving a maintained plateau with these flows.

The issue of the effect of ambient NO on FE_NO is still unsettled. Silkoff and coworkers (15) as well as Piacentini and colleagues (29) have both reported that significantly different ambient levels of NO have no effect on levels of exhaled nitric oxide, especially when nasal contamination is excluded and plateau pressures are being evaluated. We have found this to be the case. We showed that a cutoff level of 30 ppb in our initial 10 subjects did not alter our study findings. The European Respiratory Society (17) suggested that ambient levels of 40 ppb may not have an effect on FE_NO values. Since we found no differences in FE_NO values at ambient levels of 30 or less, we elected to include all subjects in our regression analyses. However, we elected to eliminate the possibility of ambient NO as a confounding variable in our ongoing studies by using a scrubber to obtain NO-free air.

Our data show a threefold increase in FE_NO values between the highest and lowest flows. This is less than the 35-fold increase reported by Silkoff and coworkers (15); however, it is expected because their range flows (4.2 to 1,550 ml/s) were considerably higher than ours (4 to 46 ml/s). In their study, values at high flows were too close to the detection limits of the analyzer and hence would introduce large sampling errors and may not enable discrimination of normal from disease states. In addition, these higher flows would be difficult to sustain, especially in younger children. In fact, in our laboratory younger children (less than 8 yr old) had some difficulty in sustaining the flow of 46 ml/s. Franklin and colleagues (18), using mouth pressures of 10, 15, and 20 cm H₂O, corresponding to flows of 50, 75, and 100 ml/s, found levels of NO (10.3, 8.5, and 7.4 ppb, respectively) lower than ours, but that would be expected for their higher flows. In our experience, the higher flows may not be practical for children under 8 yr of age although these authors included 7- to 13-yr-olds in this study (mean age 9.7 yr). At higher flow rates, children may achieve full exhalation before a plateau is achieved.

Recognition of the significance of flow is important in children because the relative flow in children will vary widely depending on lung size. An optimal flow would be low enough to amplify the NO signal, to avoid measurements near the analyzer detection limits, to allow a better distinction to be made between normal subjects and those with disease states, to achieve plateau in a reasonable time, and be easy to maintain for younger children. The suggested flow should also be practical over a wide age range in children. At this time, it seems premature to limit the flow to one absolute rate. However, on the basis of our experience, a flow rate of 30-50 ml/s in children seems to satisfy these requirements, with 50 ml/s as the most well tolerated.

In our model flow, body surface area, age, and FEF_25-75 were significant predictors of FE_NO. Silkoff and coworkers reported a regression model using flow as the only predictor of FE_NO in adults (15). Franklin and colleagues found a 12% increase in FE_NO with each year of age, but found no influence of age, FVC, and height (18). Their data and ours suggest that other factors may play an important role in children who are still growing. Our preliminary model will need to be confirmed in future studies. We feel that the use of our technique (including the four flow rates used in this study) will enable us to determine the optimal flow to be used in children of all ages, as well as provide reference values for comparison between laboratories.

We chose expiratory pressures of 10 and 20 mm Hg in order to generate the flow rates used in our studies. These pressures are not necessarily needed because it is likely that the lowest pressure that reliably closes the velum is acceptable for FE_NO measurements as long as the flow is determined. This contention is supported by Silkoff and coworkers (15), who demonstrated similar values for exhaled nitric oxide with the application of 60 mm Hg mouth pressure and those obtained with 20 mm Hg pressure. This finding would support the contention that NO measured is predominantly airway in origin and of a lesser degree than delivered to the alveolus through the vasculature (30). Children in our study complained of no discomfort and noticed no differences when different pressures were used.

There are three main approaches to measuring NO: the constant flow, single-breath technique; the tidal breathing method; and, the single-breath reservoir technique. Each of these techniques has its own appeal and may be useful in specific situations. Our method, the constant flow single-breath technique, requires cooperation and has been the method of choice in adults. However, we and others have demonstrated that this method may be used on children who are able to cooperate (10, 18). In fact, the use of this method may enable us to determine relative flows in relationship to lung size and hence may allow comparison from various laboratories and across age groups. The disadvantage of this method, as pointed out by Canady and colleagues (33), is the fact that the analyzers are neither inexpensive nor portable, which makes their widespread use impractical. These authors contend that use of reservoir bags containing mixed vital capacity expirate may be collected in various clinics and transported to a central laboratory, and hence should be recommended for clinical convenience. They reported that online and vital capacity measurements were linearly related (R² = 0.88), which would suggest that this method may have clinical utility and should be studied further.

Our bag collection techniques were slightly different from that used by Canady and coworkers (33) in that ours consisted of collection of the remaining exhaled gas under similar conditions as online analysis, while their subjects exhaled against resistance into a free bag. Therefore it is not surprising for the most part that the exhalation gas we obtained resulted in FE_NO values similar to that obtained online. However, the offline (bag) method collects deadspace and washout phase of the exhaled gas. This would explain findings using our flow rates, since at higher flows, there will be less discrepancy between online and offline values as the differences between the online plateau and the NO in the deadspace and washout phase is less.

The tidal breathing method and the single-breath reservoir technique are easy to perform even in younger children, who may have limited ability to cooperate. A major disadvantage of these techniques was possible contamination with nasal NO, and the inability to determine and control expiratory flow rates. Nasal contribution has been effectively eliminated by applying a resistance in the expiratory line to keep the soft palate closed (15, 16, 34, 35). However, the inability to control expiratory flows may render these methods less useful in comparison of data from one laboratory with those of another.

Although the numbers in each group were small, we found a curious increase in FE_NO values from 15- to 17-yr-olds, and a decrease in 18-yr-olds. This occurred across all flows measured. The increase in FE_NO with age has been found by others; however, their upper age limit was 13 yr (18). Dinarevic and coworkers found FE_NO values in adults as high as in children, which may be a reflection of the relatively higher flow rates in children when similar techniques are used (36). Our 15- to 18-yr-olds confirm an increase in values until age 17 yr, followed by a decrease at age 18 yr. The reason for this finding is unclear, but is unlikely to be due to changes in lung volume with age. A more plausible explanation may be hormonal or other unrecognized physiological changes.

While other investigators have found no consistent effect of age, sex, or height on exhaled nitric oxide levels (22, 37, 38), we found lower levels of FE_NO values in females as compared with males at the four higher flows. There is evidence that whole-body production of NO, as judged by an increase in urinary nitrites and nitrates, is greater in premenopausal women (between Days 7 and 14 of the menstrual cycle) than in men (39). However, the role of NO is controversial because increased (40, 41) or diminished (42, 43) production in women as compared with men has been reported. Whether this finding is due to an increase in production of FE_NO in males or to a decrease in females in our study group is unknown and requires further study.

In summary, we have attempted to further define the expiratory flow rates that may be appropriate in children. Our studies would suggest that the ideal flow rate for children may be in the 30- to 50-ml/s range. In addition, FE_NO values may also depend on other factors such as age, body surface area, and lung function tests. Our methodology (i.e., using four flows in the range suggested) is a useful starting point for testing of children at all age groups because the likely optimal flow would be in this range. In addition, the reference values obtained would enable us to compare FE_NO values obtained from various laboratories and may lead to further understanding of the role of FE_NO in airway inflammatory disease and the role of antiinflammatory therapy. The age-related and sex- related changes we have found are difficult to explain and warrant further study.