Arthur A. Meltzer MPH; Michael H Sm...

Arthur A. Meltzer MPH; Michael H Smolensky PhD; Gilbert E D'Alonzo, terminate F.C.C.P.; Ronald B. Harrist, PhD; and Peter H Scott MD

We examined the relationship across 24 hours between percent-predicted values (PPV) of peak expiratory result (PEF) and forced expiratory book in one second ([FEV.sub.1]) in a arrange of 23 stable untreated asthmatic children 6 to 17 years of age from means of regression analysis as well as the percentage difference between the PPV of these pair measurements. Although the Pearson correlation coefficient between the PPV was consistently high, ranging between 0854 and 0892 the assumption that like a finding substantiates the substitution of PEF for [FEVsub1] is called into question. through 50 percent of the make subordinates displayed a 10 percent or greater difference in the PPV between the sum of two units measurements, regardless of the time of day the sum of two units respiratory variable were determined, while through the whole extent of one-third of all subjects evidenced a 20 percent or greater discrepancy between the PPV of the couple measures. While, on a dispose basis, there was no statistically significant difference in the means percentage difference throughout 24 hours between the PPV of [FEVsub1] when compared with the corresponding measurement of PEF reliance forward PEF alone in individual make liables may result in a false impression of the patency of the airways in comparison to the [FEVsub1]

Measurement of PEF has been advocated for use by dint of asthmatic patients to self-monitor airway patency and assist the physician in determining the weight of therapy throughout the 24-hour period. In addition, it has been used as a tool for monitoring respiratory status among healthy and asthmatic subdues in epidemiologic field studies.[1-13] Since early findings showed a high correlation between PEF and the [FEVsub1][14-16] and because of the relative ease and lack of outlay with which the PEF maneuver may be carried revealed the former measurement has been used as a surrogate for the more sensitive measurement of [FEVsub1][17] despite admonitions against in the same state [i]or[/i] condition a practice.[18-22]

Respiratory mechanics and dynamics can vary from one side of to the other 24 hours as a follow of both voluntary changes in performance of lung function trials as well as circadian rhythmic changes in lung mechanics.[23,24] Since outpatient use of the peak follow meter may occur at all times of the day and night, it is of interest to determine whether PEF is strenuously reflective of changes in [FEVsub1] during different spans of the day and night as well as through the whole extent of the entire 24-hour period. Therefore, we compared the sum of two units measures of airway patency at specific times during the day and night in a assign places to of stable asthmatic children.

MATERIALS AND METHODS

The [FEVsub1] and PEF were measured by the agency of a Collins 9-L respirometer and mini-Wright peak pour meter, respectively, in 23 diurnally active, stable-untreated asthmatic children, aged 6 to 17 years. subdues had received no oral bronchodilator medications for at least six hours (48 hours in the case of theophylline) prior to studious mood commencement. Airway status was monitored at approximately three-hour intervals in a controll hospital setting through the whole extent of a continuous 24-hour period. The data were gathered during a placebo-control period during the investigation of circadian rhymes in airway patency.[25] All exposes were well-trained in the use of the mini-Wright peak arise meter and in spirometric conducts Three successive forced expiratory maneuvers were first done to obtain [FEVsub1] through spirometry, followed by three in the same state [i]or[/i] condition additional maneuvers to obtain PEF according to the mini-Wright peak flow meter At each clocktime assessment, the best of three measurements for the pair [FEV.sub.1] and PEF was recorded. The data were transformed to PPV using standard equations to enable comparisons to be made which controll for the variables of age, sex height, and weight.[26] In order to better evaluate tendency s in airway function over the 24 hours, the data were assemblageed into four six-hour quadrants and the mean value (x) of each subject's PPV of [FEVsub1] and PPV of PEF was obtained for each of the four time closes A separate regression equation, with hanging variable [FEV.sub.1] and independent variable PEF as well as the Pearson work moment correlation coefficient, was evolveed for each time quadrant.

The use of correlation or regression techniques to estimate the predictability of common method by another has been questioned.[27] The absolute fact that two measurements, in this case [FEVsub1] and PEF are highly correlated with single in kind another in no way implies that the sum of two units are measuring exactly the same aspects of airway status. The high correlation coefficient between the couple indicates only that [FEV.sub.1] and PEF are earnestly associated with one another. The specific nature of that association (ie, the reach to which [FEV.sub.1] and PEF measure the same aspects of airway patency) is unknown. Therefore, as an additional means of examining the relationship between the pair measurements of airway patency, the percentage difference between the percent-predicted [FEVsub1] and PEF [D = 100 x (PEF-[FEVsub1])/[FEVsub1]] was calculated for each expose for data collected at approximately three-hour intervals through the whole extent of the entire 24 hours. The mean percentage difference between the sum of two units respiratory variables over the entire 24 hours was also calculated for each subdue Each observation of D at time t was then transformed to rectangular coordinates (xy) through the equations x = DCO (2 [unkeyable]t/24) and y = Dsin (2 [unkeyable]t/24). For each of the 23 subdues [unkeyable] and [unkeyable] were calculated to create 23 pairs of the form (xy) with each pair representing the mean vector above 24 hours for the difference between the PPV of PEF and [FEVsub1] for that make subordinate Each vector has the couple a magnitude, indicating the amount of the difference between the brace variables, as well as a direction indicating the time (on a 24-hour clock) at which the average difference occurr These pairs were assembled in a 23 x 2 matrix and Hotelling's T[2 28] was used to standard the hypothesis that the mean of the bivariate observations is naught (ie, there is no significant discrepancy in the percent predicted difference of PEF and [FEVsub1]) The 95 percent confidence ellipse for the differences was also calculated.