INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Since the effect of rise time to peak expiratory flow (PEF) was first suggested as possibly affecting the performance of PEF meters (1), it has been important to determine how the shape of the flow-time profiles in the region of PEF varies in the normal and the diseased populations. The American Thoracic Society has recently proposed a new range of human flow-time profiles specifically for testing PEF meters, and these profiles include a range of rise times to PEF (2) (see Table 1). We have recently published the range of rise and dwell times found among normal subjects when recording their PEF with a pneumotachograph (3), but data from patients with airflow limitation have not yet been presented. Patients with severe chronic obstructive pulmonary disease often have a very brief ill-sustained peak on their flow-volume curves, and so it seems likely that the dynamic characteristics of the way they achieve their PEF may be different from those of normal subjects.

We have therefore undertaken a retrospective analysis of blows recorded from patients attending a lung function laboratory in order to establish the 95% confidence limits for these indices in patients.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

All patients who underwent testing in the lung function laboratory of Good Hope NHS Trust Hospital (Sutton Coldfield) had their maximal forced expiratory maneuvers recorded using an unheated Fleisch pneumotachograph (PT) whose performance has been previously described (4). The PT was placed on a fan between blows to maintain thermal stability (5). The analogue output from the differential pressure transducer (FC040; Furness Controls, Bexhill, UK) was passed through a 100-Hz low-pass filter (Butterworth 6 pole) and sampled at 250 Hz, with the flow, time, and integrated volume signals being stored on computer disk whenever the incremental volume exceeded 20 ml or elapsed time exceeded 50 ms, whichever occurred the sooner. The PT system was calibrated on a daily basis by discharging a 3-L syringe through the PT several times using a different range of flows each time. The calibration factor was calculated by equating the mean integrated volume from these discharged blows to 3 Litres (6). The range of flows ensured that there was no bias in the calibration to a certain flow range. All recordings were corrected to BTPS and also for the viscosity differences between the calibrating gas and expired gas using the formula of Sutherland (7) and Wilke (8).

The frequency response of the PT system was tested using a closed low frequency loudspeaker system and calibrated microphone. The amplitude response of the PT was tested in the range 2 to 50 Hz and compared with the recording from the microphone. The response was flat to 30 Hz, which was deemed more than satisfactory for recording the maximal forced expiratory maneuver (9). The adequacy of the PT frequency response was verified by testing with a profile from an explosive decompression device (10) that can deliver profiles with a very rapid rise to PEF (3).

The recorded blows for all patients from 1989 to 1994 who were older than 17 yr of age were checked, and subjects were included only if their blows fulfilled BTS/ARTP recommendations for acceptance (11), with the additional criterion that the time from 0.1 L/s flow (the flow threshold to start the recording) to peak flow should be less than 300 ms. From these blows, for each subject, the blow with the largest peak expiratory flow (PEF) was selected for further analysis using a computer program that calculated the time taken for flow to rise from 10% of the PEF value to 90% of the PEF value, and this was called the 10 to 90% rise time (RT). The elapsed time with flow equal to or in excess of 90% of the recorded PEF was found by interpolation and this was called the 90% dwell time (DT90). The derivation of DT90 is presented graphically in Figure 1, and a similar process was undertaken for 95% of PEF (DT95). The subject's PEF, FEV₁, FVC were calculated from the selected blow and related to predicted values using the ECSC equations (12) and the method of standardized residuals (12, 13) in order to remove age, height, and sex bias.

View larger version (22K): [in this window] [in a new window]   Figure 1.   The first 0.3 seconds of a flow time profile plotted, with the derivation of the 10 to 90% rise time (RT) and the 90% dwell time (DT90) outlined.

Statistical analysis was undertaken using SPSS version 6 for Windows, and a level of 5% was taken as significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The PT system was tested using the explosive decompression device with a profile that produced a flow of 11.30 L/s with an RT of 38 ms and a DT95 of 15 ms, which the PT system recorded as a PEF of 11.38 L/s with an RT of 39 ms and a DT95 of 17 ms. We conclude that the PT could satisfactorily record human flow profiles with sharp rise times and short dwell times.

The blows from a total of 912 patients older than 17 yr of age were available for analysis. Of these there were 556 men and 356 women. The mean, SD, and range of values for their spirometric data are shown in Table 2. As these subjects were all patients, their spirometric indices when expressed as standardized residuals (SR) were on average below predicted, and this was most marked for FEV₁ and FEV₁%, indicating that patients with airflow limitation were more common than patients with restrictive disease, upper airway obstruction, or who had normal lung function. The frequency distributions for RT, DT90, and DT95 among all the subjects are shown in Figure 2. These indices were not normally distributed and showed a negative skew. The values for each index for all the men were significantly smaller than that found for all the women (p < 0.0001, Mann-Whitney test). The median and selected percentile values in milliseconds are shown in Table 3.

View larger version (31K): [in this window] [in a new window]   Figure 2.   Frequency histograms for the 10 to 90% rise time, the 90% dwell time, and the 95% dwell time for all the men and the women.

The number of patients with all their spirometric indices within the normal range, i.e., above the lower 95% confidence limit of 1.645 SR, was 153 men and 124 women, and these subjects were deemed "normal." Patients with FEV₁, PEF, and FEV₁% all below 1.645 SR numbered 220 men and 85 women, and these subjects were deemed to have airflow limitation. There were 183 men and 147 women from the total number of subjects whose data did not fall into these arbitrary categories. The mean, median, and range for the rise and dwell times for those subjects deemed normal and those with airflow limitation are shown in Table 4. When comparisons were made between the rise and dwell times for these groups of subjects it was found that for both the men and the women these indices were significantly smaller, i.e., shorter, in the subjects with airflow limitation (p < 0.0001, Mann-Whitney test). In both the normal subjects and those with airflow limitation the DT90 and DT95 were significantly shorter in the men than in the women (p < 0.005, Mann-Whitney test), but there was no significant sex difference for RT. This suggests that the sex difference in RT when all the subjects were analyzed together was because a larger proportion of the men had airflow limitation.

Spearman's rank correlation between the rise and dwell times and the PEF as absolute values and standardized residuals are shown in Table 5. Pearson's correlation coefficients were all similar and slightly smaller but still significantly different from zero correlation. In the normal subjects the RT was shorter if the PEF was larger or above predicted, the same being true for the dwell times. In the subjects with airflow limitation the RT relationship with PEF was weaker and in the opposite direction with a shorter RT if the PEF was smaller or below predicted, whereas the dwell times retained the same relationship with PEF as the normal subjects.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We have presented the first data outlining the range of rise and dwell times for PEF in patients with abnormal lung function. These data indicate that for patients with airflow limitation the rise and dwell times tend to be shorter than in normal subjects.

We have previously presented the rise and dwell times in normal healthy subjects recorded using a similar pneumotachograph, with the median RT being 88 ms for women and 70 ms for men (3). In the present study on a larger number of patients with normal spirometry results and a wider age range we found the median to be 72 ms for both men and women. However, in subjects with airflow limitation the range of values was much lower for all the indices. This was not just because the achieved PEF was lower in patients with airflow limitation since in the normal patients a larger PEF was significantly associated with a shorter RT. In the patients with airflow limitation the opposite was found, with a small RT being associated with a small PEF. Further work on how the acceleration to PEF varies with disease states may help elucidate the relationship between achieved PEF and RT. An alternative explanation would be that the force-velocity relationship in the respiratory muscles in patients with airflow limitation is very different from that in normal subjects, but we are not aware of any evidence that this is the case.

The subjects tested in this series were all patients referred to a lung function laboratory for various reasons. A large proportion of these had suspected pulmonary disease, and so they could not be classed as a randomly selected normal population. Those patients in this study with `normal' spirometric function has rise times very close to those found in our previous smaller study on normal volunteer subjects (3). Therefore, we believe that these data will be representative.

There are a number of possible applications for these data. Some of the criteria set for quality control of spirometry have been based on back-extrapolated volume and the variability of FVC (2, 11). Which rejection criteria are adopted for selecting blows as satisfactory may need to be adjusted according to the circumstances of the recordings. Increasingly patients are making recordings of their own lung function in a nonclinical setting and with no supervision. The recording of FVC requires subjects to continue until no more air can be expired, and when unsupervised this may not always be achieved. Therefore, FVC repeatability may not be a good criterion to use under these circumstances. Also rejection criteria based on back-extrapolated volume as a percent of FVC may also be less reliable when patients are unsupervised during recording. If PEF is the most important index being recorded, which is usually the case for the monitoring of asthma, then the use of the RT to PEF may prove helpful to make sure the blow was started with adequate effort. The upper 97.5th percentile limit for RT could easily be programmed into electronic handheld PEF meters and a blow rejected if the RT exceeded this limit. This limit was 150 ms in the normal male patients and 154 ms in the normal female patients, so a single figure of 150 ms might be suitable for this purpose. The use of the absolute time to PEF, measured using a back extrapolation technique has previously been investigated in the context of quality control measures in spirometry (14). These investigators found the estimated upper 95th percentile was 88 ms for men and 115 ms for women. These results can be expected to be lower than the 10 to 90% rise time values we have obtained because the absolute time to PEF was measured from a time zero defined by the back-extrapolated technique.

Another application for these data is in the construction of flow time profiles suitable for testing the performance characteristics of PEF meters. The ATS has endorsed the use of a panel of 26 flow time profiles for this purpose (2, 15). These seem to have been selected to represent the range of possible blows to be recorded. However, they do not necessarily take into account the full range of profiles that can be expected. The correct recording characteristics of a PEF meter should be that it records accurately if the rise and dwell times are both long, and the frequency response must also be sufficiently good to record the correct PEF when both the rise and dwell times are short, that is, no overshoot that would indicate underdamping and no undershoot that would indicate overdamping. From the data presented here it is possible to construct a limited panel of flow-time profiles that encompass the range of rise and dwell times found in both normal subjects and those with airflow limitation and have these scaled to cover the range of flows encountered. This approach would lead to a simpler testing procedure that is based on population data rather than on single samples from subjects. The range of RT and dwell times for the 26 ATS profiles in Table 1 do not match the ranges we have found in our study. The lowest RT in the 26 profiles is in the region of the median value for patients with airflow limitation, and the same is true for dwell times. Therefore, a limited panel of profiles based on our data might better reflect the range of profiles encountered in practice.

We conclude that the rise time and dwell time of PEF are shorter in patients with airflow limitation than in normal subjects, and this relates to the changes in lung mechanics leading to the airflow limitation and is not due to the fact that their PEF is lower.