Thorax 1991;46:643-650 643 Reference values and prediction equations for Thorax: first published as 10.1136/thx.46.9.643 on 1 September 1991. Downloaded from normal lung function in a non-smoking white urban population C Michael Roberts, Kenneth D MacRae, Andrew J Winning, Lewis Adams, W Anthony Seed Abstract ptoms,'l' whereas in others no definition is Prediction equations for normal lung given.89 Some studies have even discarded function have been derived from tests on low values arbitrarily, using minimum 179 healthy, non-smoking, white urban spirometric values as selection criteria for nor- dwellers. The subjects, 96 women (height mal subjects.' "° Most recent studies have ex- 1-461-77 m) and 83 men (height 1-61- cluded smokers" '7 because the values found 1-96 m) aged 18-86 years, underwent even in asymptomatic smokers may deviate measurements of spirometric flow and from those in non-smokers. volume, multi-breath helium dilution Some published studies have used values lung volumes, and single breath carbon for population subgroups defined by age, sex, monoxide transfer factor and the single race, or religion,7 1-21 or have produced nor- breath nitrogen washout test. Regression mal ranges for a limited number of tests analysis using height, age, and weight as only. 31416 independent variables was used to As a result, many lung function laboratories provide predicted values for both sexes. have been forced to adopt reference values Correlation coefficients were similar to derived from several different lung function those found in previous studies but nor- studies. In an attempt to overcome the lack mal ranges for spirometic measure- of comprehensive data the European ments were narrower than in many Community for Coal and Steel working party previous studies, and spirometric flow published compound prediction equations and volume measurements were higher recommended for all people of European des- than those obtained in studies that cent.22 These equations are based on data from included cigarette smokers, reflecting our studies carried out over 30 years, and thus more stringent criteria for selecting incorporate the inconsistencies that arise from http://thorax.bmj.com/ subjects and the newer standardised the use of variously defined normal groups. technical methods adopted. Multi- They also include data from smokers, and breath helium dilution values for total from studies that were not carried out accord- lung capacity were similar to those ing to currently recommended technical found in previous studies but the in- guidelines.2223 They are not weighted for the spiratory vital capacity was larger and numbers included in individual studies, and the residual volume reduced. Values for the normal ranges about the mean predicted carbon monoxide transfer factor and the values are "best estimates" with little statis- on September 30, 2021 by guest. Protected copyright. single breath nitrogen washout did not tical validity. Finally, the decline of lung fun- differ significantly from existing values. A ction with age seen in longitudinal studies has complete set of lung function reference differed significantly from the decline predicted values and prediction equations for both from cross sectional data,2427 suggesting that sexes has been derived from a single data from cross sectional studies reflect the population. The exclusion of cigarette population of that time and are valid only for a smokers and subjects with respiratory limited period before population changes ren- symptoms has produced values that der them obsolete. Until long term longitu- should have a greater sensitivity in the dinal studies are completed cross sectional detection of mild lung disease. data should be renewed at intervals. The present study was stimulated by a research requirement for contemporary normal ranges Normal lung function values and ranges are for some lung function values. The opportu- Department of conventionally calculated according to varia- nity was taken to provide a comprehensive set Medicine bles such as age, height, and weight, which of lung function prediction equations for both C M Roberts contribute independently to predictions of A J Winning sexes, based on contemporary technical and L Adams lung function. The prediction equations statistical methods with rigorous selection W A Seed should be derived from a normal and rep- procedures for normal subjects. Statistics Unit resentative population tested by standardised K D MacRae technical methods and subjected to Methods Charing Cross and appropriate statistical analysis. SUBJECTS Westminster Medical as as normal were School, London Many of the older studies still used Subjects were defined if they W6 8RF reference values fail to meet these criteria. lifelong non-smokers (not more than one Reprint requests to: The definition of normal subjects has varied cigarette a day for one year), had no history of Professor Seed widely and in some has included cigarette respiratory or other serious illnesses, and gave Accepted 28 May 1991 smokers and individuals with respiratory sym- negative replies to a simplified Medical 644 Roberts, MacRae, Winning, Adams, Seed Research Council respiratory health question- residual capacity (FRC), starting from 8% and naire, which covered current respiratory 12% helium respectively (in 20-9% oxygen, the symptoms, previous medical history, time lost balance being nitrogen). Measurements were Thorax: first published as 10.1136/thx.46.9.643 on 1 September 1991. Downloaded from from work, and smoking history. repeated on each subject with a 15 minute Subjects were mainly recruited through interval until two FRC determinations were advertisements placed in the hospital foyer within 200 ml of each other; usually two and and fracture clinic; some elderly subjects were occasionally three attempts were required. recruited as a result of advertisements placed During the rebreathing period two inspiratory in a pensioners' newsletter. Eighteen subjects vital capacity (IVC) manoeuvres were perfor- were actively recruited from among hospital med. Tidal volume breathing was recorded staff and medical students to provide a balance throughout the test, which continued for at of the extremes of height. Of the 204 subjects least seven minutes, and until the helium who volunteered for the study, 25 were rejec- concentration had remained stable for at least ted because of a history of smoking (9), cough 60 seconds. The residual volume (RV) was or dyspnoea mentioned on the questionnaire derived from the FRC minus expiratory (8), concomitant non-respiratory disease (3), reserve volume and total lung capacity (TLC) or failure to meet the volume reproducibility from RV plus the higher IVC value. The mean criteria of the lung function tests (5). of the results from the two tests was taken. Each subject was measured for height by stadiometer (Harpenden) and weighed. Age, Carbon monoxide transfer capacity sex, and occupational details were recorded. Carbon monoxide transfer factor (TLCO) was measured by the single breath method29 on APPARATUS both Gould and Morgan systems with a 10 Because of a laboratory upgrading, two second breath holding time. Carbon monoxide separate systems were used for measurements was measured by electrochemical (Gould) or of airflow, volumes, and transfer factor. Eighty infrared absorption (Morgan) cells with an six subjects were studied with the Gould lung initial concentration of 0-3% in 20 9% oxygen function computer (model CPI 5000 IV) and and 10-14% helium, the balance being 93 by a combination of the Morgan lung nitrogen. function transfer test (model C) and the At least two attempts were made and accep- Hewlett Packard Vertek pneumotachograph ted if the IVC fell within 10% of the IVC (model 5000 VR). The comparability of these obtained during the multi-breath helium dilu- systems was studied with 15 normal subjects, tion study. The TLCO from the first two accep- who were tested on both systems on two table tests was taken. Alveolar volume was different days. Intrasubject variability was measured simultaneously from the dilution of evaluated in 15 subjects tested on the Gould 10% (Gould) or 14% (Morgan) helium and http://thorax.bmj.com/ unit on two occasions on separate days within was used to calculate the transfer factor correc- one week at the same time of day. The nitro- ted for volume (TLCO/VA). The mean of two gen washout test used a combined Warren E acceptable tests was taken for the TLCO and Collins photo-optical nitrogen ionisation TLCO/VA. sensor -(model 21232) and a Hewlett Packard Vertek pneumotachograph for all subjects. Single breath nitrogen washout Before the tests the required manoeuvres This was performed as described by Buist and were demonstrated by the operator and Ross' after the other tests. Expiratory flow was subjects were encouraged and supervised displayed on an oscilloscope trace to provide on September 30, 2021 by guest. Protected copyright. throughout the performance of tests. All tests visual feedback to assist the subject to maintain were carried out between 0900 and 1800 flow rates below 0 5 1/s. An IVC within 10% of hours. Calibration checks on all instruments the IVC recorded during the multi-breath were carried out at least weekly. helium dilution test was required for a test to be accepted. Values from two tests meeting these MEASUREMENTS criteria were recorded; closing volume (CV), Expiratory flow-volume curve closing capacity (CC), and slope of phase III Expiratory flow-volume tracings were recor- were