Document 3Jjbbzedd3gKDGkGx1yErvQNn

SPECIAL COMMUNICATION . Changes in Measured Spirometric Indices* What Is Significant? Bernard E. Pennock, PbJD.;f Robert hi. Rogers, M.D., F.C.C.P.;T and D. Robert ifcCafree, M.D. //- $'-p/ PLAINTIFF'S EXHIBIT - ST00055G I D ulmonary function testing in tbe last decade has moved from the research laboratory into the daily practice of internal medicine and surgery. The first report of a simple method of assessing air ways obstruction was in the late 1940s.14 It is not surprising that in this relatively young field the in terpretation of the tests is still not universally clear. In recent discussions in the pulmonary section of'^, the University of Oklahoma at Oklahoma City, one specific question stood out In a subject with dem onstrated abnormal measured spirometric values, how much acute or chronic change is required to be certain that improvement or deterioration has oc curred? In searching the literature we concentrated on the evaluation of spirometry, since this has been examined in tbe greatest detail. The following assess es the data from the literature, as well as some of our own data, and attempts to answer the question. We recognize that this is not necessarily die final an swer, but hope it will generate some interest and comments and make us pause and think about what we label improvement and deterioration. Significance The usual definition of a clinically significant dif ference from normal (an abnormal measured value) is a measured value that statistically would be ex-\ pec ted in less than S percent of normal subjects. In ) a Caussian population, this represents values that are more than 1.65 times the coefficient of variation below the mean. For example, the coefficient of variation for FVC determined in a large group of normal subjects (Table 1) is about 13 percent Av. value 21.5 percent (1.63 X 13) smaller than the J mean is considered abnormal. ---- ^ From th Pulmonary Disease Section. Department of Med icine, University of Oklahoma Health Sciences Center, and Veterans Administration Medical Center, Oklahoma City, tPresently at Pulmonary Disease Section, Department of Medicine, University of Pitriburgh School of Medicine, Pittsburgh. Reprint requests; Dr. Rogers, Pulmonary Medicine, Seaife Hall, Unnxmfy of Pittsburgh, Pittsburgh 15261 Population Significance The application of this significance criterion to the parameters of spirometry in normal population studies is illustrated in Table L Coefficients of varia tion, using a 30-year-old, 173-cm tall man, axe given from the studies of Morris et a],' Kory et al,4 and Knudson et al.* The approximate average of the jnrfmVHih of variation for FVC. FEVi, and ^EF 25-75? are 13, 13, and-25 percent, respectively. NlulTfpTyTng these coefficients of variation by 1.65 \ will give significance levels of about 21, 21, and 41 J percent, corresponding to die commonly used cri teria that values less than SO, SO, and 60 percent of the predicted values are normal values.* Significant Chances in Time One must identify coefficients of variation for mul tiple measurements within a day (acute changes) and for determinations from week to week (chronic changes) to develop criteria for the interpretation of time sequential changes in values for individual patients. Acute Chance Three studies that have presented data with mul tiple daily spirometric values are listed in Table 2. McCarthy et alT tested 12 normal subjects ten times each day with a coefficient of variation calculated as shown in Table 2. Pennock et al* tested 20 sub jects with reversible airways obstruction nine times 1--'Population Coefficient of Voriotion* Reference (No. in Study) Morris* (517) FVC 14.5 FEV, FEF 25-75% 13.8 26 Kory* (468) 12.1 13 Knudien* (128) Summary 12J 13.7 13 13 23.2 25 `Coefficient of variation calculated on the basis of a 30-yearold, 173-ctn tall male. CHEST. 80: 1. JULY, 1981 CHANGES IN MEASURED SPIROMETRIC IKOICES 97 .. . - ------------------------------- -- < Tabi 2--(Within Day) Coefficient /. FcrttUti Reference (No. of Piljtnu, No. of Tests Each) FTC FEV, FEF 23-73 % McCarthy' (12, 10) 13 8 Pcaaock* (20, 0) 0.7 8.1 14 Hruby** (13, 10) Norma) Obstructive Restrictive 33 7 9 11 Summary Norma) Obstructive 33 6.7 8.1 8 14 *Riay, not cotflWicot of nrotiM. daily and found values for the coefficient of varia tion about twice those of McCarthy's normal sub jects. Similarly, Hruby et al,* testing IS subjects ten times per day, found that the ranges of coeffi cient of variation in both the obstructive and re strictive- subjects were about twice those of his normal subjects. Multiplying each of the average values for coefficient of variation by 1.63 yields the result that a significant change within a day's time in a normal subject is 5 percent or greater in FEVi and FVC and 13 percent in FEF 25-75?. However, insignificant change in an obstructive patient is about twice that of a normal subject, ie 11 pericent of FVC, 13 percent in the FEVi, and 23 per cent in the FEF 25-755. As a general rule of thumb. Tables 1 and 2 show that the daily individual vari ability in all tests is about one-fourth the population variability in normal subjects and about one-balf of V.the population variability in chronically obstructed ^Subjects. Chronic Chance To evaluate long-term changes, one must look at the coefficients of variation from week to week. Table }<--ZnJiWafl (Wrek-to*Wttlt) Coefficient of Variation Rtfercoc* (No. o( Patkau, No. of Tctti Each) F"VC McCarthy' (20, 10) and (5, 23) 3 Spicer" Normal (11, 30-60) Bronchitic (7, 30-60) Asthmatic (10, 30-60) 7.8 13-5 14.1 Penooclc' (20, 3) Asthmatic 11.1 Summan* Normal Obstructive 6.4 FEV, FEF 23-75% 7 13 14.2 18.4 7 13 18.4 T*bU 4--Significance Level (Chmyj* Create* than Percent Sho**n) Population main tVithin day Normal Obstructive IVcck-to-wesk Normal Obstructive FVC 21 5 11 11 21 FEV, FEF 23-75% 21 41 5 13 13 23 12 21 23 30 Three studies have been identified in which these data are available (Table 3). In normal subjects significant week-to-week changes are 11 percent in the FVC, 12 percent in the FEVi, and 21 percent in the FEF 25-75?. The patients with obstructive disease had greater variability than normal subjects, the variability being 21` percent in FVC, 23 percent ' in FEVi, and 30 percent in FEF 25-755. Again, in week-to-week changes, the subjects with obstruc tion showed about twice the variability of the nor mal subjects. Table 4 summarizes the data for population stud ies and for longitudinal studies in individuals, both normal subjects and subjects with airways obstruc tion. These levels define significance at the 95 per- -- cent confidence limit, based on the cited population studies. ST00055G2 Appucation As an example of the use of this type of informa tion, one of the tests frequently used in evaluation of obstructed patients is the measurement of spi- rometric parameters before and after administration of a broncbodilator. From Table 4 it is apparent that an increase in FVC and FEVt by more than approximately 12 percent; 11 percent in FVC; 13'v\ Jpercent in FEVj, or an increase in FEF 25-75? of ) greater than about 25 percent (23 percent) repre- sents a statistically significant change from the base line. Similarly, if one is following up a patient over a long term, to assess a stable, deteriorating, or im proving condition, the week-to-week change (Table 4) in FVC and FEVi, must be greater than 20 to\ 23 percent or the change in FEF 25-755 greater than J about 30 percent. * The identification of subjects with spirometric indices that are abnormal is commonly accom plished from a definition of clinical significance based on the concepts of statistical significance at the P<0.05 leveL We have applied these concepts to measured data to derive criteria for clinical sig nificance in relation to changes of measured spiro- 38 PENNOCK. ROGERS, UCCAFFREE CHEST, 80; 1, MY, 1981 metric value* in an individual subject within a day and from week to week. Again, we think this tjpe of analysis yield* clinically useful information, but we recognize that there-fltty be better approaches. Whether or not better approaches exist, we hope to simulate discussion of this very important question of improvement or deterioration in an individual patient Rmiuwca* 1 Tilrma R. Pinelli A. Air drculant t air captif dam rcsplomtioo da la foncMop vcntdatrfct pubnonaire. Psris Mad 1947; 133:624-28 2 Gtautlee EA. Analysis of the Mandatory defect by timed eapadty measurements. Am Rav Tobere 1951; MiSt-71 3 Mortis JF. Kosid A, Johnsoo SC Spiranttrie staadardt for healthy non-smoking adults. Am Rev Rospir Oil 1971; 103:57 4 Kory RC Callahan R, Bonn HC, Synder JC The Veter, am Adimturtreboo-Army cooperative study of pulmonary function. Am J Med 1961; 30:243 5 Knndson R/, Statin RC. Lebowa MO, Burrows B. The mammal expiratory flow.volume curve: normal standards, variability and effects of age. Am Rev Respir Dis 1976 113:587 6 Snider CL, Kory RC, L>oos HA. Cradios of pulmooary function impairment by means of puknooary function rests. Dts Chest 1967; 52:270-71 7 McCarthy OS, Craig OB, Chrmiadc RM. EEect of modifleatioo of Che smoking habit on hsng function. Am Rev Respir Ois 1976; 114:103 8 Fennodt BE, Rogers RM. An evaluation of tests used to meajuxe brooeboddatsr drug topcoat. Chart 1978; 73(mppl):988 9 Htuby J, Butler J. Variability of routine pulmonary func tion teats. Thoraa 1975; 30:548 10 Spicer 'VS, Kerr OH. Variation in respiratory function. Arch Environ Health 1966; 12:817 r ST0005563 CHEST. 80: 1, JULY, 1981 CHANGES IN MEASURED SPIR0METR1C INOICES 99