Document kJGYQRGj9JgebmdOVRLqQwRJ

sen, Maclaren If foe-, covayate W5;4i:91-I02. >ndon: Chapman ig men with early iosis. Edinburgh: IOM report No implications of sis in working conference on the i Conference of 9. (Annals of the iredict attacks of The Statistician % first three irticles or 'breviated ; and the article or nces are: >rs. Uniform nedical jour:aria; release phil chemoige. N Engl of invading , Sodeman nechanisms 1974:457-72. British Journal of Industrial Medicine 1987;44:673-681 Development, radiological zone patterns, and importance of diffuse pleural thickening in relation to occupational exposure to asbestos H BOHLIG,1 A CALAVREZOS2 From D-4047 Dormagen/Zons1 and Department of Haematology,2 St George General Hospital, D-2000 Hamburg l, Federal Republic ofGermany abstract The radiographic appearance of the lateral pleura was divided into an upper, a middle, and a lower zone. Bilateral changes of the pulmonary layer of the pleura (diffuse pleural thickening) within the upper pleural zones were found in 863 (71%) of 1204 workers exposed to asbestos and in 249 (40%) of 622 non-exposed controls. Downwards along the chest wall this ratio of 7:4 increased progressively up to 10:1 at the lower parts of the pleura. Bilateral diffuse pleural thick ening in at least two adjacent zones on each side was found in 652 (54%) of exposed and in only 86 (14%) of unexposed subjects. The difference was even more striking when comparing bilateral involvement of all three zones (28% and 3% respectively). Unilateral change was rare (4-8% and 7-8% respectively) and often due to causes other than exposure to asbestos. Pleural findings were the earliest radiographic features detectable associated with former exposure to asbestos. Bilateral diffuse thickening in at least two adjacent zones on each side seems to be a striking feature and an early indication of former occupational asbestos damage. Modifications of the International Labour Organisation 1980 classification are proposed. Diffuse pleural thickening after exposure to asbestos has been well known to the pathologist for nearly 100 years.12 The clinical evaluation of this feature started much later, with the advance of radiology,3 7 but a broader knowledge of its frequency and significance did not come until the introduction of the high kilo volt technique for investigating the chest.1 8 Although international publications covering this topic have increased tremendously, many diagnostic ians show little or only a superficial interest in this common feature of the objectively recordable sequelae of exposure to asbestos.9'20 None the less, since 1968 the International Labour Organisation (ILO) in Geneva has developed a semiquantitative recording of pleural changes and added it to the Inter national Classification of Pneumoconioses.21 22 This pleural part of the scheme provides recording facili, ties by special numbers and letters for diffuse pleural ,. thickening for the side (R/L), the length (0, 1, 2, 3, Accepted 10 November 1986 measured in quarters of the chestwall), and the width (a, b, c, measured in millimeters) of the pleural shad ows. There is little information, however, on the validity of this recording system. In 1983 in a preliminary investigation we analysed whether these facilities were efficient enough to correlate--with some degree of reliability--pleural changes with asbestos and other fibre dust exposures. Evaluation of 1031 German workers exposed to asbestos indicated that nearly all hitherto existing codes of the formula "RLla" and many of "RL2a" were epidemiologically irrelevant.23 The present study describes pleural changes in radiographs of 1204 individuals with an occupational exposure to asbestos from the medical surveillance in 1982 and compares them (retrospectively and in a non-blind manner) with films of 622 randomly selected controls with no known asbestos history. Several variables are included in the analysis. The purpose of this study was to identify distribution pat terns of diffuse pleural thickening after exposure to asbestos and to give proposals for a revision of the pleural part of the ILO .1980 classification.22 673 HWBUI0001460 674 Table 1 Main characteristics ofgroups A and B Bohlig, Calavrezos Group A: No (%) Group B: No {%) Significance ofdifference A :B Women Age: Range -20 -- 30 -40 -50 -60 >60 Ethnic origin: German Turkish Years since first exposure to asbestos: Median -5 -10 -15 -20 -25 -30 >30 Thorax deformity: Barrel shaped Bell shaped Others Overweight* ILO: emphysema Former thoracic traumaf 1093 (90*8) 111 (9-2) 42 16-70 12 (1-0) 198(16-4) 295 (24-5) 438 (36-4) 233(19-4) 28 (2-3) 1033(94-1) 71 (5-9) 10 303 (25-1) 359 (29-8) 213(17-7) 118 (9-8) 91 (7-6) 71 (5-9) 49 (4-0) 98 (8-1) 91 (7-6) 12 (1-0) 231(19-3) 15 (1-2) 33 (2-7) 382(61-4) 240 (38*6) 39 15-73 73(11-7) 127 (20-4) 132(21-2) 126 (20*3) 100(16-1) 64(10-3) 577 (92-8) 45 (7-2) -- -- -- -- -- -- 75(12-1) 53 (8*5) 6 (1-0) 97(15-6) 9 (1-4) 36 (5-8) p < 0-001 x p < 0-001 NS 3 NS| NS I NS! NS || ^Exceeding 30% more than the Broca index. tOr any reason for pleural thickening other than asbestos exposure. (Adjusted for age. Adjusted for sex. NS |] Not significant, adjusted for age and sex. Materials and methods POPULATION Radiographic films of two groups were chosen: a sur vey group of unselected individuals occupationally exposed to asbestos from different geographic regions (group A) and a non-exposed control group (group B). The main characteristics of the groups are presented in table 1. Group A consisted of 1204 workers undergoing medical examination in 1982 with a history of occupational exposure to asbestos for at least one year. They had been exposed to nearly all types of asbestos dust generated within the German asbestos processing and handling industry, including construc tion industries, but there were no shipyard workers. Working conditions, duration of exposure (minimum one year), and dust concentrations could not be assessed, because the information available was inad equate. In general, individuals with less than 15 years since first exposure may be representative of plants with lower dust concentrations owing to an effective medical supervision complying with legal require ments established during this time. Those with a longer history were probably exposed to much higher dust concentrations and not so homogeneously con trolled. Initially, group A consisted of more than 1300 consecutive cases, but the number decreased to 1204 after eliminating those with poor film quality (= "2," "3," and "4" for recording technical quality). Group B consisted of 622 controls with no known occupational exposure to asbestos: 202 were first films of workers entering the asbestos industry with no pre vious exposure to asbestos. In addition, 420 radio graphs of individuals most of whom were generally engaged in an industrial environment and a few of their clinical data were randomly selected from the archives of two large community hospitals (St George General Hospital, Hamburg, and Municipal Hospi tal, Luedenseheid). Films of this type and provenance were excluded when clinical information (or the films themselves) indicated pleural changes attributable to acute diseases such as pneumonia or pleurisy. Owing to the retrospective design of this study, the films were not homogeneous in quality, exposure time, or positioning. In particular, those of group A originated from more than 90 sources (B readers). METHODS OF INVESTIGATION To avoid interobserver variation, which exists for pleural changes in an even wider range than for parenchymal findings.24 all films were reread by one reader (HB) who is one of the four qualified A readers in the Federal Republic of Germany and who has been familiar with recording pleural changes for nearly four decades and who--in addition--has been HWBUI0001461 'ig, Calavrezos difference A:B Development, radiological zone patterns, and importance of diffuse pleural thickening 675 for age. i uality {= "2," uality). ith no known were first films y with no prem, 420 radiovere generally and a few of pted from the Jals (St George fiicipal Hospiid provenance p (or the films Jttributable to eurisy. (his study, the iity, exposure ;e of group A B readers). ch exists for ige than for eread by one red A readers ind who has ! changes for bn--has been Fig l Chest x ray (a), zones (b}, and diffuse pleural thickening as used in this study, (b)follows schemefor lung zones and records diffuse pleural thickening compared with (a) representing different length ofaffected chest wall according to six zones. This offers unique opportunity ofalso recording localisation (zone pattern). Pleural caps (arrows) are taken as "0" = NAD when restricted to second rib (L), whereas in case ofrunning down chest wall below clavicle upper zones were recorded as affected--that is, positive (R). a member of the ILO group of experts designing the last two versions of the ILO classification.21 22 For the purpose of this study, diffuse pleural thick ening has been defined as the measurable pleural com panion shadows seen in profile along the bony chest wall on straight frontal projection (solely posteroanterior films). En face findings were not recorded, since they are easily misinterpreted. For the recording of diffuse pleural thickening, the ILO scheme and the modified German system were adapted to allow for additional information on extent (length), side, and localisation. Thus the six zone scheme, as already used for recording lung appearances, was used. The information about length is given automatically in terms of affected zones: upper, middle, and lower zones (UZ, MZ, LZ) and their combinations (fig 1). Following the ILO instructions for discriminating the three lung zones of each side, the length of the bony chest wall of the lower zone is considerably greater than the length of the upper zone. As may be seen in fig 1, this modification leads therefore to a recording system of length with four different grades of extent, but one should be aware that the length of pleural thickening indicated by zones is not necessarily the same as in the original ILO grades of extent (0, I, 2, 3 (see ILO22)). In the upper zones symmetrical apical pleural thickening of up to 5 mm of width was recorded as not affected (fig 1, left side) whereas pleu ral thickening that extended below the level of the clavicles (fig 1, right side) was recorded as affected. Other than posteroanterior films, no additional views, especially obliques--widely discussed in other publications10 12 25--were available to us. On the other hand, according also to our experience, obliques may increase the risk of wrong positive records of pleural thickening due to subpleural fat,25 especially in the regions of the posterior parascapular and axillary lines as shown in many computed tomo grams of the chest. The extent and width of pleural thickening and most parameters of the ILO 1971 and 1980 classification were recorded. In addition, the variables overweight (exceeding 30% more than the Brocaindex), thoracic shape (barrel shaped, bell shaped, scoliosis), or thoracic deformities due to trauma or surgery, the time variables "period since first exposure" and "age at examination" were included in the analysis. Since it seemed possible that in the sub group of Turkish workers some individuals might have been previously exposed to an endemic fibrous dust,26 27 they were investigated separately. Turkish individuals in both groups were younger than in the German subpopulation (only three older than 50). In the exposed Turkish workers the period since first exposure was correspondingly shorter. STATISTICAL ANALYSIS Descriptive statistical methods and tests for unadjusted homogeneity of proportions were used as described in the SPSS.X statistical package.28 In com parative analyses of proportions and life time data,29 adjustment for confounding variables was done by using combined risk estimates30 or maximum like lihood estimates (MLE) of the common odds ratios of 676 combined tables29 31 based on algorithms modified from Thomas and Gart29 and Thomas32 and by using multivariate logistic models as proposed by Breslow and Day.33 Usually, the unconditional tests have been applied. With sparse data, conditional MLE methods were additionally used to confirm results of unconditional tests. The probability distributions of risk curves were estimated by the classic method of Kaplan and Meier.34 Owing to inequalities in the distributions of age and sex in groups A and B, adjustments for these vari ables were regularly performed. This is not mentioned further in the text. Since pleural thickening after exposure to asbestos is known to be time dependent, analyses adjusted for time are essential. By contrast with a prospective cohort study, no time intervals were available for group B that were comparable with the period since first exposure of group A. For an ad hoc substitute, age at examination was taken as the time variable for time adjusted comparison of groups A and B serving as a supplement to the unadjusted analyses of proportions. Results Unilateral pleural thickening was rare (table 2) and often associated with other causes of pleural thick ening such as trauma (table 1). Therefore only the results of bilateral symmetrical involvement are reported. Table 2 also shows highly significant differ ences in typical asbestos sequelae such as plaques or pulmonary fibrosis (small opacities of a profusion of 1/0 or more), whereas obliteration of the costophrenic angle (cpa) as a possible indication for a so called asbestos pleuritis17 is equally rare in both groups. Table 3 shows zone patterns of bilateral pleural thickening in groups A and B. Relation to age at Bohlig, Calavrezos Table 2 Case numbers of unilateral diffuse pleural thickening and ofsome other items of the ILO 1980 classification in both groups Significance Group A Group B of (n = 1204} (n = 622) difference p Unilateral pleural thickening Obliteration of costophrenic angle (cpa = +) Circumscribed pleural thickening (plaques) Profusion ofsmall opacities (1/0 and more) 58 24 215 167 49 15 4 7 NS l| NS|| p < 0-001 p < 0-001 j| See footnote to table 1. examination is shown in table 4. Table 5 presents the relation between zone patterns in group A and the period since first exposure. The risk estimates of developing bilateral pleural thickening in relation to time (period since first exposure for group A and age at examination for comparing group A and B) are collected in figs 2 and 3 a-f. All these data show that bilateral pleural thick ening was significantly (throughout p < 0 0001) more prevalent in individuals exposed to asbestos than in those unexposed. After a period since first exposure of 25 years the individuals in group A had a risk of bilat eral complete (UZ, MZ, and LZ) pleural thickening of 50% (fig 2). These 25 years correspond roughly to a median age at examination of 54. At this age, only 5-2% of the individuals in group B had developed extended bilateral pleural thickening (fig 3f). By con trast. pleural thickening of the upper zones often involved individuals of the control group (fig 3a). One quarter of that group showed isolated bilateral upper zone involvement, and in another 14% it was com bined with other zones. This represents a relation of Table 3 Bilateral diffuse pleural thickening of upper, middle, and tower -ones in both groups Zones Upper 0 + 0 0 4+ 0 4- Middle 0 0 4- 0 + 0 44- Lower 0 0 0 + 0 444- 00 0 + + 0 4- 0 + 0 -1- +0 4- 0 4- 0 4- 0. pleural thickening absent; +. present; 0. absent or present. Group A No ( % ) 223 (18 5) 224 (18-6) 99 (8-2) 5 (04) 296 (24-6) 1 to 1) 14 (1-2) 342 (28-4) 1204(100) 223 (18-5) 863 (71-7) 751 (624) 362 (30*1) Group B No (% > 358 (57-6) 166 (26-7) 7 (11) 5 (0-8) 67 (10-81 -- 3 (0 5) 16 12-61 622(100) 358 (57 6) 249 (40-0) 93 (15 0) 24 (39) HWBUI0001463 f, Calavrezos ,eurafi 1980 ; Significance { of | } difference p NSJ NS|| p < 0 001 p < 0 001 5 presents the p A and the estimates of in relation to up A and age V and B) are ileural thick0-0001) more estos than in : ;t exposure of i risk of bilatal thickening id roughly to this age, only id developed l 3f). By con- zones often i (fig 3a). One lateral upper it was coma relation of p B No (%) (57 6) (26-7) (11) (0-8) (10-8) ; (0-5) f (2-6) j!00) j(57-6) ! (40-0) 1(15-0) i (3-9) Development, radiological zone patterns, and importance of diffuse pleural thickening Table 4 Zone pattern and age. both groups u ML A/B No Age (years) -20 AjB (%} -30 AjB (%) -40 . AjB (%) 00 +0 0+ 00 ++ +0 0 4+ 4- No + +0 + 0 4+ 0 +0 0 0 0 4* 0 44+ 4-0 4-0 + 223/358 224/166 99/ 7 5/ 5 296/ 67 1/ -- 14/ 3 342/ 16 1204/622 863/249 751/ 93 362/ 24 50/70 33/27 --17//---- --/ 3 ---/--- / - /- 12/73 33/27 17/ 3 --/" 46/ 68 28/ 27 4/ 1 19/ 4 --/ -- -/ - 4/ 1 198/127 51/ 32 26/ 6 4/ 1 18/ 52 26/ 34 8/ 1 1/ 1 27/ 7 01 -- 01 l 19/ 5 295/132 72/ 45 54/ 13 20/ 6 0, bilateral pleural thickening absent; 4-, present; 4- 0, present or absent. U, upper; M. middle: L, lower. -50 AjB (%> 12/ 54 15/ 27 9/ 2 0/ 2 27/ 13 --/ -- 2/ 35/ 3 438/126 77/ 43 73/ 18 37/ 5 -60 AjB (%) 7/.51 9/ 16 10/ 2 0/ 1 23/ 25 --/ -- 2/ 2 49/ 3 233/100 81/ 44 84/ 32 51/ 6 677 >60 AjB (%) 11/52 4/27 14/ 2 --/ 2 25/16 --/-- 7/ -- 39/ 3 28/64 68/45 86/20 46/ 5 Table 5 Zone pattern andperiod sincefirst exposure to asbestos, group A Zones No u M L UM UL ML All Z Period (years) -5 No/% -10 No/% 100/ 33-0 91/ 30-0 22/ 7-3 1/ 0-4 56/ 18-5 --/ -- 2/ 0-7 31/ 10-2 303(100) 84/ 23-4 72/ 20-1 24/ 6-1 f/ 0-3 89/ 24-8 __j __ 6/ 1-7 83/ 23-1 359(100) -15 No/% 24/ 11-3 38/ 17-8 18/ 8-5 --/ -- 9_6j/ _32-4 2/ 0-9 62/ 29-1 213(100) -20 No/% 9/ 7-6 13/ 110 11/ 9-3 2/ 1-7 35/ 29-7 --/ -- 1/ 0-8 47/ 39*8 118(100) -25 No/% 4/ 4-4 6/ 6-6 12/ 13-2 ---/ --23/25-3 i/ i-i i/ i-i 44/ 48-4 91 (100) -30 No/% 1/ 14 3/ 4-2 9/ 12*7 --j -- 13/ 18 3 ---/ --2/ 2-8 43/ 60-6 71 (100) >30 No/% 9/ 3*0 1/ 3-0 3/ 6*7 1/ 3*0 11/ 24*4 --/ -- --/ -- 32/ 71*1 49(100) No 223 224 99 5 296 1 14 342 1204 Period since first exposure (years) Fig 2 Estimated risk ofbilateral pleural thickening in different single or adjacent zones irrespective ofinvolvement or no involvement ofother zones. Pleural thickening at lower zones appears in median between 7 and 12years later than pleural thickening of the upper or middle zones. As pleural thickening was recordedafter and not at first appearance value ofzero percent at time zero is probably not true. group A to B of roughly 7:4, whereas the lower zones show a relation of 10: l. Bilateral pleural thickening of the lower zones within group A developed rather late (in median 10 years later, fig 2). But it was more dis criminating than thickening of the more cranial parts of the pleura (figs 3c, e, and f). The female breasts sometimes give problems in analysing pleural shadows. Within the German sub groups A and B, no significant difference between men and women was found. This makes a reading bias improbable. In the Turkish subgroups bilateral involvement of all six zones of group B (mainly in men) occurred nearly as often as in group A (p = 0-73). In addition, the Turkish subgroup A developed this feature somewhat earlier than the German subgroup A. Chest abnormalities named above or emphysema ("em") were of no significant influence, although the bell shaped thorax may cause some difficulties in diag nosing middle and lower zone involvement. With a borderline statistical significance (p = 0-04), a slightly HWBUI0001464 678 higher prevalence of "bilateral pleural thickening" was observed in overweight individuals of group A. This was expected and may be due to subpleural fat. Bohlig, Calavrezos It was of interest to determine whether the age at onset of exposure in group A influenced the develop ment of bilateral extended pleural thickening. From Bilateral pleural thickening ot middle zones 100 (5) o Exposed *o Not exposed 80' p<00001 /*60 o 40 CL 20 / Bilateral pleural thickening ot lower zones 100 o Exposed Not exposed .o .o' 80-| pO-OOOl 160 -Q S 40 20- 0 -r- 10 20 30 40 50 60 Age at examination Bilateral pleural thickening ot upper and middle zones 100 (d) o Exposed Not exposed 80 p <0-0001 /* 60- i !1 40- 20- 0- to 20 30 40 50 Age at examination 60 Bilateral pleural thickening ot middle and lower zones 100-j|(e) o Exposed ,o Not exposed 80- p<00001 t 60- ao 40 a. 20 0 0 10 20 30 40 50 60 Age at examination Bilateral pleural thickening ot all zones 100- o Exposed /'C ' Not exposed 803 p<0-0001 p'' t6o- j5 ao 40a_ 20- 0 ^8----*-- 10 20 30 40 50 60 Age at examination 0- --'O -o--o- ----*-- 10 20 30 40 50 60 Age at examination Fig 3 a-f Estimated risk ofbilateral pleural thickening in different single or adjacent zones in relation to age at examination irrespective ofinvolvement or no involvement ofother zones. Pleural thickening of upper zones is less specific and that oflower zones, which is almost always combined with thickening at the adjacent middle zones, is most specific for asbestos exposure, . tv pleural thickening was recorded after and not at first appearance value ofzero per cent at age IS is probably not true. 'dig, Calavrezos ther the age at 'ed the developickening. From Idle zones o--o T i-------- 1-------- 1------- 1 50 60 id middle zones ! ,o-- r' / r 50 ' 60 zones /.o .o' ) ---------!__ ! 50 60 id examination id that oflower m exposure. As jt true. Development, radiological zone patterns, and importance of diffuse pleural thickening 679 Period since first exposure (years) Fig 4 Estimated risk ofbilateral pleural thickening at different ages at time offirst exposure (p < 0-0001. after adjustmentfor period sincefirst exposure: p < 0 0001). First exposure at older age has a higher risk ofbilateralpleural thickening ofall zones. As pleural thickening was recorded after and not atfirst appearance value ofzero per cent at time zero is probably not true. fig 4 it is apparent that the risk increases with age. This difference was also found after adjusting for different observation times (period since first exposure was--as could be expected--in median six years longer in the age group up to 20 years than in the group over 40). Discussion Our results show that bilateral diffuse pleural thick ening is the most frequent6 14 24 25 and probably the earliest detectable sequel of exposure to asbestos with the current strong legal restrictions for the asbestos industry. In the exposed group there was not a single case of parenchymal involvement that did not also have concomitant, usually advanced, diffuse pleural thickening. This may be true only for the recent con ditions of low level dust exposure. Therefore, we find the assessment of pleural involvement to be of strik ing epidemiological importance in the medical superveillance of dust exposed populations. In this context it is noteworthy that pleural thickening may also occur in association with other types of dust exposure.8 Nevertheless, as shown above, such diffuse pleural lesions commonly appear in unexposed popu lations, being more prevalent at higher ages and especially in the upper zones. On the other hand, the more impressive feature of pleural plaques is extremely rare in the unexposed group B (see table 2). Thus bilateral pleural thickening per se. although present in a much higher prevalence than plaques, dis criminates exposed subjects only from a statistical point of view but not on an individual basis. This is an important difference. As this study emphasises, however, additional information on the localisation (affected zones) per mits a more specific differentiation as to whether a given pleural thickening is due to dust exposure or not: bilateral involvement of at least two adjacent zones was a characteristic feature of exposure and lower zones usually become affected later than the other parts of the pleura. Unilateral pleural thick ening is, by contrast, of limited value in relation to asbestos exposure, although some association may exist to former pleuritis due to asbestos or not. Also, the involvement of the upper zones only is less significant, since this feature was found in a quarter of the unexposed population. This invalidates nearly all the readings of the codes RLla or RL2a or both (see ILO21 2Z). In summary, this implies that the middle zones are the most important regions for the pleural sequelae of asbestos inhalation. Accordingly, the absence of bilateral pleural thickening in the middle zones with involvement of other zones instead was a rare feature in the exposed group and should there fore always suggest causes other than asbestos. This importance of middle zone involvement should be strongly considered in case the ILO "Recommen dations for future research" for recording in quarters instead of sixths of the lungs should become accepted. The high prevalence of pleural thickening of the upper parts of the pulmonary pleural layer in the exposed population is in striking opposition to the behaviour of asbestosis in the lung. Likewise, the rather lower and late involvement of the lower zones by contrast with the middle zones is equally sur prising. An explanation of this different predilection for reactions of asbestos dust in the lung and the pleura is not apparent from this study. It is knownthat shorter fibres and more isometric dust particles gain the upper bronchi, whereas the longer fibres reach rather the lower parts of the lung, favouring asbestos changes in these locations.1 8 35 36 Appar ently, the short fibres are predominantly transported to the pleura through the lymph vessels. Thus the selection of the fibrous dust by the airstream may explain the described phenomena. The differences in the reported prevalence of pleu ral thickening are highly influenced by national and international interobserver variation and the personal biases of the readers. Former international reading trials have shown that each reader has his own char acteristic "handwriting" when recording pleural findings. For that reason reports on prevalence are difficult to compare. For our study, the films of both groups, A and B. were reread by the same reader. The range of specific personal characteristics must there fore be the same in both groups and makes the results HWBUI0001466 680 Bohlig, Calavrezos at least comparable between A and B, even though a possible bias cannot be excluded with certainty as the reader was not "blind" to asbestos exposure. Never theless, there is a remarkably higher prevalence of bilateral pleural thickening in our study than in other investigations, as for instance in that of Cordier et al.13 On the other hand, there is no doubt that training and personal exchange of knowledge and experience among readers can decrease interobserver variation. In reading pleural findings the agreement (identical records) among the four A-Readers in the Federal Republic of Germany in 1978-9 was--owing to such exchange--high (87-90% of 14000 films.)2* Despite some other insufficiencies of this investigation (retro spective design, no blind reading, limited number of available controls, and only vague data on occupational conditions) the striking differences between groups A and B are conclusive. With respect to the high proportion of Turkish workers within the working population of the Federal Republic of Germany, the knowledge about endemic pleural changes after exposure to fibre dust in Turkey26 27 made a separate analysis of individuals with Turkish ethnic origin necessary, since our files showed no exact data on the regions where they came from. The higher prevalence of diffuse pleural thick ening within both Turkish subgroups A and B may be at least partially due to this precondition. Conclusions As figs 2-4 show, the development of diffuse pleural thickening is first obvious within the upper parts of the pleura and not in the lower zones (except after pleurisy); the speed ofdevelopment is much slower in the unexposed group. This discriminates exposed and unexposed populations from a statistical point of view but not on an individual basis. The main feature of discrimination is the bilateral involvement of at least two adjacent zones; the "mandatory" involve ment of the middle zones rarely occurs in nonexposed individuals. This type of involvement requires recognition and inclusion in the pleural part of the ILO 1980 scheme in terms of pleurally affected lung zones, as already in use for recording lung appearances. The study presented shows that this method is useful and more convenient than the present recording system of extent of length and side, since it offers the hitherto lacking, but necessary, indication of localisation and therefore more reliable information. This seems par ticularly important since, as was shown under low level dust exposure, diffuse pleural thickening is the earliest and thus a sensitive radiological sign of the. biological action of asbestos. References 1 Otto H, Bohlig H. Morphologic und Rontgenologic der Ashesrose. Radiologe 1985;25:9-21. 2 Stewart MJ. Pulmonary asbestosis. Br Med J 1928;ii:675-82. 3 Bohlig H, Jacob G, Muller H. Die Asbestose der Lungen-Genese. Klinik, Rontgenologie. Stuttgart: G Thieme. 1964. 4 Jacob G, Bohlig H. Die ronigenologischen Komplikationen der Lungenasbestose. Fortschr Rontgenstr 1985;83:515-25. 5 Kiviluoto R. Pleural calcification as a roentgenological sign if nonoccupational endemic anthophyllite-asbestosis. Ada Radiol I960;suppl 194. 6 Kiviluoto R, Bohlig H. Die Pleura. In: Bohlig H, ed. Die Slaublungenerkrankungen und ihre Differentialdiagnose. Stuttgart: G Thieme, 1964;261-7. 7 Sluis-Cremer GK, Theron CP. Radiological and pathological correlations in asbestosis in the Republic of South Africa and the United Kingdom. In: Whipple HE, ed. Biological effects of asbestos. Ann NY Acad Sci 1965;132:373-8. 8 Bohlig H, Otto H. Die pleuralen Rontgenzeichen der Pneumokoniosen. In: Handbuch der medtzinische Radiolagie. Vol lX/5b. Berlin: (in press). 9 Abeida SM, Epstein DM, Gefter WB, Miller WT. Pleural thickening: its significance and relationship to asbestos dust exposure. Am Rev Respir Dis 1985;25:9-21. 10 Baker EL, Greene R. Incremental value of oblique chest radio graphs in the diagnosis of asbestos-induced pleural disease. Am J Ind Med 1982;3:17-22. 11 Begin R, Boctor M, Bergeron D, et al. Radiographic assessment of pleuropulmonary disease in asbestos workers: posteroanterior, four view films and computed tomograms of the thorax. Br J Ind Med 1984;41:373-83. 12 Bohlig H, ed. Staubtungenerkrankungen und ihre Differentialdiagnose. Stuttgart: G Thieme 1964. 13 Cordier S, Theriault G, Provencher S. Radiographic changes in a group of chrysolite miners and milters exposed to low asbestos dust concentrations. Br J Ind Med 1984;41:384-8. 14 Hillerdal G. Non-malignant asbestos pleural disease. Thorax 1981;36:669-75. 15 Mattison MC. Asbestos and asbestos related disease. Croydon: Jupiter Press, 1983. 16 McGavin CR, Sheers G. Diffuse pleural thickening in asbestos workers: disability and lung function abnormalities. Thorax 1984;39:604-7. 17 McLoud T, Woods BO, Carrington CB, Epler GR, Gaensler EA. Diffuse pleural thickening in an asbestos exposed population: prevalence and causes. Am J Roentgenol 1985;144:9-13, 18 Wagner JC, ed. Biological effects of mineral fibres. Lyon: Inter national Agency for Research on Cancer, 1980. (tARC sci publ No 30.) 19 Wain SL, Roggli VL, Foster WL. Parietal pleural plaques, asbes tos bodies, and neoplasia. Chest 1984;85:707-13. 20 Whipple HE. Biological effects of asbestos. Ann NY Acad Sci 1965;132:373-8. 21 International Labour Organisation. ILO U!C international classification of radiographs of pneumoconioses 1971. Geneva: ILO, 1972. (Occupational safety and health series 22 (rev).) 22 International Labour Organisation. Guidelinesfor the use of ILO international classification of radiographs of pneumoconioses. Geneva: ILO, 1981. (Occupational safety and health series (rev 80).) 23 Bohlig H. Calavrezos A. Localization of pleural thickening and its correlation to asbestos dust exposure. In: Bergbau Berufsgenossenschaft, ed. VI International pneumoconiosis confer ence. Bochum. 1983. Vol 2. Bremerhaven: Wirtschaftsverlag NW. Verlag fur neue Wissenschaft, 1984:890-8. 24 Bohlig H. Interbeurteiler-Variation bei arbeitsmedizinischen Vorsorgeuntersuchungen von Asbestarbeitern in der Bun desfepub1ik"Deutscljland. In: Bergbau-Berufsgenossenschaft, ed. VI International pneumoc oniosis conference. Bochum 1983. HWBUI0001467 ?, Calavrezos 1| J ; >logie der Asbes- l8;H:6T5-8Z Lungen-Genese. 4 964. 4 nplikationen der | 3:515-25. ' nological sign if T-sbestosis. Acta \ 11. ed. Die Stau~ | -gnose. Stuttgart: ! | md pathological South Africa and p ologica! effects of - \en der PneumoRadiologie. Vol ler WT. Pleural J to asbestos dust tque chest radioeural disease. Am ] aphic assessment ; workers: postomograms of the ? ihre Differential- \ phic changes in a d to low asbestos 84-8. disease. Thorax Hsease. Croydon: :ning in asbestos finalities. Thorax R, Gaensler EA. (>sed population: 5;144:9-I3. 'res. Lyon: Inter9. (IARC sci publ al plaques, asbes-13. nn NY Acad Sci r/C international es 1971. Geneva: cries 22 (rev).) for the use of ILO f pneumoconioses. hnd health series kl thickening and In: Bergbau Berhoconiosis conferIwirtschaftsverlag h8- leitsmedizinischen |rn in der BunIfsgenossenschaft, ce. Bochum 1983. Development, radiological zone patterns, and importance of diffuse pleural thickening 681 Vol 3. Bremerhaven: Wirtschaftsverlag NW, Verlag fur neue Wissenschaft, 1984:2096-103. 25 Sargent EN, Boswell WD, Ralls WP. Markovitz A. Subpleural fat pads in patients exposed to asbestos: distinction from noncalcified pleural plaques. Radiology 1984;152:273-7. 26 Baris YI, Sahin AA, Ozesmi M, et al. An outbreak of pleural mesothelioma am! chronic fibrosing pleurisy in the village of Karain/Urgup in Anatolia. Thorax 1978;33:181-92. 27 Saracci R. Epidemiology of groups exposed to other mineral fibres. In: Wagner JC. cd. Biological effects of mineralfibres. Vol 2. Lyon: International Agency for Research on Cancer, 1980:951-61. (IARC sci publ No 30.) 28 Nie NH, Hull GH, Jenkins JG, Steinbrenner K, Bent D. Statisti cal package for the social sciences. New York: McGraw Hill, 1975. 29 Thomas DG, Gart JJ. Stratified trend and homogeneity of proportions and life time data. Comput Biomed Res 1983;16:116-26. 30 Mantel N. Haenzel W. Statistical aspects of the analysis of data from retrospective studies of disease. / Natl Cancer Inst 1959;22:719-48. 31 Gart JJ. Statistical analyses of the relative risk. Environ Health Perspect 1979;32:157-67. 32 Thomas DG. Exact and asymptotic methods for the combining of 2x2 tables. Comput Biomed Res 1975;8:423-46. 33 Breslow NE, Day NE. Statistical methods in cancer research. Vol 1. The analysis of case control studies. Lyon: International Agency for the Research of Cancer, 1980. 34 Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. Journal of the American Statistical Association 1958;53:457-81. 35 Doll R, Peto J. Asbestos. Effects on health ofexposure to asbestos. London: HMSO, 1985. 36 Elmes PC. Health hazards of short mineral fibers. In: Proceedings ofthe symposium on "Skortand thin mineralfibres," Stockholm, 1982. Stockholm: Government of Sweden, 1983:163-80. Destruction of manuscripts From 1 July 1985 articles submitted For publication will not be returned. Authors whose papers are rejected will be advised of the decision and the manuscripts will be kept under security for three months to deal with any inquiries and then destroyed. HWBUI0001468