Document 2jaxe2vd4bV3OvpvExXQGmm5L

JUL-13-2001 FRI 04^ 09 PM FAX NO, P, 12 THE LANCET 11 SacrenjR. Gillon S, Fecgwon A. Incidence of inflammaro.-y he#*; disease tn Scottish children between 1968 and 1993: marginal fa# sa ulcerative cciitil, three-fold rise in Crohn1! disease. Gut 1989; JO: 613-22. 12 Miller C, Feirington CP, Herbert K_ Tlie cpidcmiulogy ofeubsenn ideKiinipiticncaphalttii in England and Wajee 1970-1089, faf} lipi&miol 1992; 21: 998-1006. 13 Stavatw Ciij Toy PT, Taylor PE, tee T, Yip H-Y, Prospects for casmti ofhaparitis B virus infection; implications of childhood vaccination end ',e ng-term protection. Ptduuriti 1992, Wh ]7ft-7>, 14 Statitiici of infectious diseases. London: OPCS Serin MS2 no 2, 197s. 15 Gilat T, Hacohen D, Lilos P, Longman MJS. Childhood factata in. ulcerative colitis and Crohn's disease. StandJ Qaa&atlaol 198"; 22: 1009-24. 16 Wsictclman JL Lyles CM, Sandler R. Childhood infections and the risk of inflammatory bowel disease. Dir Ou Sd 3 994; 39:535-40, 17 Griffcn DE, Ward BJ. Diflarenlinl CD4 T cell activation in (neurits, 3 tt&tI Du 1993) 168: 275-81. 18 Hall AJ, Cutis FT. Unions from measles vaccina,inn in developing countries. BMJ 1993; 307: 1294-95. Is lung cancer associated with asbestos exposure when there are no small opacities on the chest radiograph? Paul Wilkinson, David M Mansell, Judy Janssens, Michael Rubens, Robin M Rudd, Anthony Newman Taylor, Corbett McDonald Summary This study was designed to test the hypothesis that the risk of lung cancer from asbestos exposure is confined to persons with radiographic evidence of pulmonary fibrosis. Occupational and smoking histories were obtained from 271 patients with a confirmed diagnosis of primary lung cancer and 678 referents (279 with other respiratory disease and 39 with cardiac disease). Histories were reviewed blind to assess the timing, duration, and probability of exposure to asbestos, To allow for a lag between asbestos exposure and the development of Song cancer, subjects were classified by the time they had spent In an occupation entailing definite or probable exposure more then IB years before diagnosis. The presence and extent of fibrosis was assessed blindly from chest radiographs by three readers and scored for small opacities with the 110 1989 International Classification of Radiographs of the Pneumoconioses. 83 (34-3%) cases had worked in an occupation with definite or probable asbestos exposure compared with 176 (25-8%) referents (crude odds ratio for lung cancer L-49, 95% Cl 1-O9-2-04). After adjustmom for age, sex. smoking history, and ares of referral, the Odds ratio (95% Cl) was 2-03 (1-0C-4-13; in the subgroup of 211 with a median ILO score for small parenchymal opacities of I/O or more, and 1-56 (1-02-2-39} In the 738 with e score of 0/1 or less (ie, those without radiologicar evidence of pulmonary fibrosis}. These results suggest that asbestos is associated with lung cancer even in the absence of radiologlcatty apparent pulmonary fibrosis. Lancet 1995; 346:1074-78 Introduction Asbestos exposure increases the risk of tong cancer, but it is unclear whether such exposure decs so in the absence of radiological evidence of pulmonary fibrosis. In the UK and several other countries, compensation for asbestosrelated disease recognises lung cancer as asbestos related only if there is radiological or pathological evidence of lung fibrosis or diffuse pleura) thickening. In other countries, compensation is based on duration of exposure in relevant occupations. Cohort studies show that the excess of lung cancer in asbestos-exposed workers is at least as great ,as the number of cases of mesothelioma.' Do Vos Irvine and colleagues estimated that 5-7% of lung cancers in Glasgow and the west of Scotland were asbestos related.3 this is in line with estimates from the USA.' There are around 35 000 new lung cancers each year in England and Wales,* of which 2000 or so may be asbestos related. Data from the SWORD surveillance scheme* suggest that comparatively few such cases arc recognised, possibly because once the patient is identified as a tobacco smoker, the physician does not inquire into asbestos exposure. But another factor may be the prevailing view that pulmonary fibrosis is a necessary prerequisite. Small opacities on the chest radiograph are a useful, if imperfect, marker of pulmonary fibrosis. They can be objectively assessed with the ILO 1989 International Classification of Radiographs of Pneumoconioses.4 With this system, a score of 1/0 or more for small, characteristically irregular opacities is usually considered evidence of fibrosis, whereas a score of Q/l or less indicates no radiological evidence offibrosis. We used this score in a hospital-based case-referent study to examine whether pulmonary fibrosis is a prerequisite of asbestosrelated lung cancer. Linton Chest Hospital, London (P Wilkinson mrcv, i Janssens sen, m Rubens men, R M Rudd frcp); am Royal Brampton Hospital and National Heart and Lung Institute, Unden SW3 6LV, UK <D M Hanseil men, PrOf A Newman Taylor frcp, Prof C McDonald frcp) ' ' CeirsfpetttoiKw to: Prof Corbett McDonald Patients and methods Subjects were selected from routine adult admission* to the London Chest Hospital between September, 1992, and March, 1903. Cases were consecutive patients with a histologically or cytologicaily confirmed diagnosis of primary bronchial eaieinoma; referents were consecutive patients with respiratory diseases other than lung cancer and a random sample (averaging one in three) of admissions with cardiac disease (table 1). A patient was excluded if he or she had first been admitted with the .ifm . JUL-13-2001 FRI 04:09 PM FAX HO, the Lancet II 96 (35*4%) Jims cancer cases, 106 (38-0%) respirator? referent*, and 92 (23-1%) cardiac referents 'were women (table 3). Long cancer cases were on average older and bad smoked more heavily. 56% of cases came from the local population; the proportion was higher for respiratory referent* <71%) than cardiac referents (41%)- A high proportion of male cases (51*4%) had worked in a jab with smite degree of asbestos exposure, as had 35*6% of male referents. Very few women had any exposure. Them was little difference between lung cancer cases and respiratory referents in their ILO scores for small opacities* but few cardiac referents had a profusion score of 1/0 or above. 2 of the 4 respiratory referents with profusion scores of 3/2 or above had cryptogenic fibrosing alveolitis, 1 had pulmonary tuberculosis, and 1 had sarcoidosis. Agreement about ILO scores was unanimous hi 563 (76*3%) of the 738 patients with a median score of 0/1 mr less (no radiological evidence of pulmonary fibrosis) (table 4). Overall agreement was similar to that found in previous Studies in which chest radiographs have been scored by muJtiple readers with the ILO system.** In die study group as a whole, 93 (34*3%) cases had worked in an occupation with definite or probable asbestos exposure compared with 176 (25*8%) referents (OR 1 49, 95% a 1*09~2`04), Table 5 shows the result* of a logistic model which adjusts for age, sat, smoking, and origin of the patient (local or tertiary referral), and Includes exposure categorised by duration. Type of referral had only a small effect on the association between asbestos exposure and lung cancer, and led to a realistically insignificant improvement in model fit. However, it was considered a potential confounding factor, and was retained in the model. Bach of the other variables had A larger influence on the observed association between asbestos and Jung cancer, and led to significant improvement in model fit. The logistic model showed that ORs for lung cancer increased with pack-years of smoking and with age, and that patients who had worked in a job with definite or HmKwi MW 0 1 2 3 S63I3S-3, 47(60' 13 aw; 210*2 teniw riMkmd by oh# tiMtfy(rt) 1G6 (17*5) n3 ai.fl) is i2*0) 2 (0*2) Onittttftf categories (<} 3(0-9) P lMi 0 Seoesii nraM ttirveciffewit ijrtepris* (X) O 14(1*6} 0 Q Table 4; Agreement between the 3 readers on ILO scare far small opacities ta 949 radiographs probable asbestos exposure more than 15 years before diagnosis had a higher risk of lung cancer than those who had not. With a simple classification of exposure--any relevant job against none--the OR for lung cancer adjusted for age, sex, smoking, and type of referral was 1*66 (1*15-2*38). The corresponding OR with respiratory referents Only was 2*02 (1*26~3<24), and with cardiac referents only 1*44 (0-96-2*16). In the subgroup of 211 with a median ILO score of 1/0 or more, the OR was 2*03 (1'00-*4*13), and in the 738 with A score of 0/1 or less, it was 1*56 {;'02-2-39), suggesting an association between asbestos exposure and lung cancer in the absence of fibrosis. The corresponding ORs from separate analyses of the referent groups were also similar: LOS (M 3-3*45) for respiratory referents, and 1*41 (0*88-2*26) for cardiac referents. In the subgroup of 655 patients with a median ILO score of 0/0 or less (those for whom at least two readers agreed about the absence of small opacities) the adjusted OR for lung cancer was 1*48 (9*93-2*35); for respiratory referents, 2*09 (1*12-3*89); for cardiac referents, 1*35 (0*82-2*22); and 1*44 (0*89-2*32) when the 37 patients with any form of pleural thickening were also excluded. The results for the separate referent groups, including several not tabulated here, were taken as sufficiently .similar to justify combining respiratory and cardiac referents for the main Analysis. Tests for exposure-response are shown in table 6. The point estimates of the OR were not higher in the category of less than 5 years' exposure, although the Cite were wide. There was evidence that ORs increased with duration of exposure both in patients wife ILO scores of skstKMtse * <50 60-55 60-63 97$ . Paik**m tfetaamt* Mouse N&ib <15 15-28 3444 *45 ter* lotsusl Nets 4 5-f' *18 * Mtt*gaMhHiiNii* e t. i- * .3 . Mn ewes (nl75| 24(13-7) 36 (20-6) |4| 46(26-3) 3(1*11 19(10*9) 42(34*0) 44(26*1) 66(36*3] iM(sa*9) BS148-61 23(13-11 16(10-8) 48(274) 116(67-4) 46 (27*4) 2(6-1) 0 RWpif$*?y 66(37-3) *4(19*7) 40(23*1) 43(24-3) 41(23*7) 25 (14-5) 31(17*9) 34 (l6*T) 42(249} 121(639) 111 (4A--9i 15(8-75 6(3*5) 31(17*9} 121(83*5) ' 39(22-5) - 11(64) . 2(19) Tab:* 3: Patient*' characteristics CafOec sit!Olsin307| 75(24-4} 51(29-6) U2(36-} ,28(3-5) 67(219) 45(14*7) 60(26*1) 66(21*5) 40 (IMS) 119(36*6) 186(61-2) 34(11*1) 19(6-2) 66(21-5) 260(64*7) 43(149) 4(1*3) 0 AlemM Cwos (n-Mj 13(13-5) 11(11-5) 33 (34-4) 39 (40-6) A* 12 (12*6) 6[6*2] 29 (30*2) 26(39*2) 22(224) 49(61*0) 93 (96*0) 1(1*01 1 (1*0) 1 (1*0) 80(63*3) 15(15*6) 1(1*0) 0 flowrarcty controls (t>=106) Cifdise eant/$& {n*92) 33(31*11 24(22*6) 29 (27*4) 20(18-9) 17(13-*} 21(22-8} 2$i3S'i! 18(19*6) 50(47*2) 10 (15*11 .16(15-11 12(11*3] 12(12*3) 79(73*6) .10*1 (93-1) 1 (0*9l 1(0-9) 0 39(42-4) IS (17-4) 22(33-9} 0(9*6) 6 (6-52) 46(30-0) 88(86-7) 0 2(2*2) 2.(11) 74(69*6} 23(21*7) 7(6*8) 3(1-9) 85(92-4} ' 8(3-8] (1*1) 0 JUL-13-2001 FR1 04:09 PM * raegeaiwsfnijTl} Noiwmsgseii Sooanauseaji Largo rfii; " - McnecsrcliiHra Smalt cUI - AlvesSar Mixoif/tis&tafM&ee RMffe*!*hf WOtKOS |B*373) Crjwiciir*t,sdiss Brershiaetass OOtirustrve deep awes* Tu&fetf&ste sestfraMryffttetfsiis tj&y? disassa Mlseeiiwsou* W>er . .. 24 99 . - ' 94' ' ' ' 99 58 3 14 73 21 a; 33 18 14 99 Citrii*totsbris (s=9) ISshatmja heart disease vatvula seas disease ansOuTt O'stviSansBs other 347 27 11 14 Table 1: Hiatetogieal type of Jung cancers and diagnosis of cwfrjote relevant disease sw than 9 months before, or if no chest radiograph was available. 4 patients were excluded because their original referral letters suggested they had been referred because of possible asbestos exposure, 23 because they did not wish to be interviewed or would not complete the interview, 29 because the chest radiograph was technically unsatisfactory nr unavailable, and ? because full case notes could nor be traced. Diagnosis and reason for and source of referral were obtained fcom case notes, and occupational and smoking histories by interview outing admission by a research nurse who was not informed of die patients diagnosis and did not know use study hypothesis, Since bias might b* introduced by informing patients of die purpose of tire research, the hospital ethics committee agreed that Subjects could only be asked whether they would be willing to participate in a study of occupational factors in illness. Patients were asked to hat, with darn, every job since leaving school ofmore than 6 months' duration, the osme of employers, to describe the work, and whether the work had entailed exposure to dust tom wood, copper, asbestos, silica, or cement. At the end of the study a list, based on a previously published study of mesothelioma,'' Was drawn up of thirty-one occupation* possibly associated With asbestos exposure. Occupational histories were blindly reviewed and each job placed in one of these occupational categories or classified ns entailing no asbestos exposure. The categories were subsequently grouped by their probability of exposure (table 2). Information on smoking habit included an estimate of the average daily cigarette consumption during each decade of life, the sum of which gave the total packyears of consumption, The presence and extent ofsmall opacities and pleural changes were assessed from the cheat radiograph taken nearest to the tune of first diagnosis. Three readers independently scored each film with the ILO system.1' As nearly all cases and many controls had radiological changes which suggested the disease for which they had been admitted to hospital, each radiograph was first reviewed by a radiologist not participating in the scoring who blacked out changes such as a tumour mass or lymphadenopaihy with an opaque silhouette over fee affected quadrant; no more than two quadrants per film were obscured- 6 ousel and 9 referents were rejected because fee film was of poor quality oc because diagnostic features could not be disguised by covering two quadrants. Films were preceded in hutches of fifty and within each batch cash quadrant was covered with similar frequency in films from ewes and referents, to achieve a similar distribution of blacketi-out areas in fee two groups, even if there was no abnormality, tinder the HO system, a profusion score for smalt opacities gives by reference to a set of standard films. Scores complies two numbers, The first is fee reader's best assessment of the appropriate category, the second is whichever adjacent category had alio to be considered. They form the following 12* m no, THEIANCiT ksbriSUf $fatboris* eapotar* 51WttWrigttWi.. ' .. SsSsfte upesiH* _ AsSsaattaytfeusfoR JfMtfogtWIdes ' ' ' ' liasSatfoafeeHs ' . No of cam (*) Hast iMtmtefK) 1(04) 13(48) 2(0-71 . 5(0-7) . . 35(5-2) 6(09) . ftehriXearpaeor* femsBwSatt 8t?l8atwk fsarify -reps Fiwta/] (itsam locomotive) Raissfsgjteftter Shat-rusisl wa$w Sr;pvro/ttoc. woof si (18b) 7(2.6) 7(2-6) 1(0-4) 3[M) 9(1-1) 21(7-8) 88(19-0) W(2-2) 16(24) 4(0.6) 4 (s-e) 7(1* 41(8.1} FSijftk aifSfrit Bttar fimsum Lath* Operator Saws) caemsnsnte/mauitwis, ro engi/mcr OH/shernfcai tsrifiefy wk Print factory Pw leisorv Ralttrsek wjinaef Seaman Trapspeft waver 2(0-7) 4(1-5) 16(50) 6 {22} 9(M) 2(0-7) 9(M) 2(0-7) 24(8-9] 17(6-3) (id) 3(04) .48(7-1) 10(1-5) 9(1-31 5(0-1) 8(1-2) 3(0-4) 48(8-8) 48(6*8) ifem^*wrilufr Carps-nr-i (sot eft Nielng dies) ONil atgtneerlftg Bfyctoasl^t SeetftcaJ or^reeiird Oeotwies ftssmdsetufe Oarasc iwnsnic KaciS^ijanautgwepiim fiu&ber/wsstee rotoufeeupo SerspfsttaiwerK SiXvsyw/sft* wpervirot' Triejsneracrii'ris sogiftow , 6(2-2) 3 (HI 1(0-4) 1 (04) 1(0-4) 7 12-6) 2(0-7) 14(5-2) 3 U'l) 0 210-7) 14(21} 12(1-89 1(0-2) 1(0-2) 2(0-3) 28(4-1) 9(1-3) 19(3.0) 10(1-6) 8(0-7} 3(0*4) Onbjow sfieh'etattsS 15 rears or more bribes inosM are shown; 125 {4S45U aw ss 35r (sa-ysf) reforms had not been empoyea In my of those eeewanionas groom n Uawiewre pefloa. Inaivifeal workers may hove contributed to mere that) on orevetalofiri group. Mebn number or relevant Jobs tr perron west)-85 In coses end 076 n referent*, wtth range in both groups of 0 to S. Table 2; hrobabhlty of afcbfetto* exposure In occupational groups point scale of increasing profusion: 0/-, 0/0, 0/1, 1/0, l/l, i/2, 2/1,3/2,2/3,3/2, 3/3, Vf, Job* were divided into those associated wife definite or probable asbestos exposure, and those wife possihic, unlikely, or definitely no exposure. Patients who had worked in Jobs with definite or probable exposure were further subdivided by the length of time in these jobs: 0 years, leas thsm 5 years, 5 to 9 years, 10 years or more, Been use of the lag between exposure and fee development of cancer, only time spent in jobs more than 15 years before diagnosis was considered relevant, Crude and adjusted odds ratios for lung cancer were calculated for all subjects and subgroups defined by fee median XLO score fur small opacities. The main study hypothesis related to the group wife s median score of 0/1 or less since these patients may be considered to have no radiological evidence of fibrosis. Although email opacities may b* due to pathological changes Other than fibrosis, their absence suggests that fee patient is unlikely to have relevant fibrosis. Respiratory and Cardiac referent* were combined for fee main analyses, bur analyses were also carried out separately for ch set of referents. Their results were broadly similar- The Influence of potential confounding factors (sex, agegroup. pack-years of smoking, and type of referral) on fee disease-exposure relation was assessed by multivariate logistic modelling. Improvements in model fit were assessed from the likelihood ratio statistic, although confounding variables were included in fee model because of feeir observed or potential Influence on tht association between asbestos exposure and lung cancer; [hay did sot need to meet specified criteria for Improvement in model fit- AN quoted likelihood ratio statistic* are those obtained by fitting fee relevant variable last. Odd* ratios (OR) are shown wife >5% CIs. . Vol 345 * Apm 2S, 1995 1075 JUL-13-2001 FRI 04509 PM FfflSJO. __ THE iAN'C^r ftwwWeiW ,Hwf OR SftWtU SIS Cl tits # P 9f* M*n ' Woeisa *55 i 29* 3-23 1'53-3.25 . 17.7 1 *0495. agKwmtiwt) tSO 50-33 . eoo >?G ntegtnatf cSprMttWMKhc NOftfr <15 1S-29 30-44 *4S m SIT 319 195 211 128 329 tea m 1 0-96 0.57-14O 19* 5 <D*Q01 1-49 093-2.39 237 143-393 i 187 139.7-94 90-2 4 <0001 7-67 4-40-1+2 5-92 4.T0-1S'9 120 S'27-22-8 Hffml UKai Tenisry Ytoet**pw* 0 <5 5.9 *1,0 SIS 433 S30 74 ts nr 1 VM a-79-2'7S 0-28 1 oet 4 i-sa 090-2.79 951 3 0037 222 MS-4-29 1*2 0.97-3.37 HfiSUMhpcORitie gsMsus. dfcdigrew of fratfem. Taefe S; HMvft* of legjatle model 0/} or less and in those with scores of 0/0, Only 8 women had any relevant asbestos exposure. Lung cancer was rare below the age of 40, and few younger patients had sufficient time to accumulate significant exposure given the 15-yetff lag. Construction and electrical work contributed most to excess of asbestos exposure in lung cancer cases, although insulation and foundry work were also important, Patients wt inaccurate in assessing their asbestos exposure! only 121 (44-9%) of the 269 patients classified as asbestos exposed on the basis of their occupation reported that they had worked with asbestos, and 42 (15-5%) reported exposure, although their jobs suggested they had none, la the subgroup with ILO scores of 0/1 or less, the adjusted OR for self-reported exposure was 1-19 (0-73-1*91), which was lower than the estimate based on objective assessment ' One participating physician in the study (RMR) had a special interest in asbestos-related disease, but exclusion of patients first admitted under his care made little difference to the point estimate (OR in the subgroup with ILO scores of 0/1 or less was J *61 [0-86-2-99]). Of the 127 patients occupationally exposed less than 15 years before first diagnosis, all but 20 bad also been exposed more than IS years before diagnosis. The independent effect of recent exposure could not therefore be determined with precision but, in the study group overall, exposure within 15 years of diagnosis and not before bad an OR of Ml (0-34-3-59). This point 7*4/* mtxfStiii* UO*n14 <1/0 0- <0 >10 ILOxsftiO/O 0 <10 *10 *VbX4 (.P-73S) i -S'9$ iO'47-2'06) S-SSfl.24-S.S0) 1*55 (0-81-3*5) (frrfSS) 1 0-83 (0-37.IBS) 2-4211*14-5.13) 14S {OSS-2'83) M*n 2**0 yn [1,-449) i 0-90 [0-41-1-98) 3.33(1.51-7.32) 1.48 (0-84-254) (f*39J) i 0-78 (0-3J.I-TBJ 3'23 (1*4-7-43) 1+2 (0-78-2-84) Table 6: Adjusted OR* tor lung cancer In relation to radiological score and years of excoscre estimate is lower than the corresponding figure for pre-lag axpoaumfOR J-66 l:15-2-38J). ' Discussion ' :* Our results do not suggest that radiological evidence'of fibrosis it a prerequisite for asbestos-related lung cancer, ho patients without parenchymal opacities, there was a 56% excess of lung cancer in those who had a high probability of occupational asbestos exposure at least 15 years before with some, albeit weak, evidence of a relation with duration of exposure. Two Issues affect interpretation of our findings; the adequacy of occupational histories in reflecting asbestos exposure, and the sensitivity of chest radiography in detecting pulmonary fibrosis. The occupational classification we used discriminated well between mesothelioma cases and controls lh a survey in the USA and Canada,' and we have no reason to think that it would not have done so adequately for lung cancer in our study. The sensitivity of the chest radiograph for detecting pulmonary fibrosis is about 80% of cases with histological evidence.mi! However, sensitivity estimates will depend on the specificity and sensitivity achieved by radiologists and histologists, and selection of cases. Staples and colleagues have shown that high-resolution computed tomography may reveal abnormalities in around a third of asbestos- exposed workers without small opacities on the chest radiograph.11 This method may thus be a more Sensitive indicator of asbestos-related fibrosis than radiography. In this, as in most case-referent studies, it is impossible to eliminate all bias. But patient interviews and exposure assessments were carried out without knowledge of the underlying diagnosis, and we believe that the use of opaque silhouettes on the chest radiograph was successM in disgusting the underlying diagnosis, although on some occasions an area of fibrosis might have beat obscured. Although patients may be selective in their recall of asbestos-exposed jobs, the systematic way in which jobs were recorded and classified, and the fact that patients themselves were inaccurate at assessing their own exposure, make it unlikely that this was a significant source of bias. Since allocating jobs to particular occupational categories was carried out blind to diagnosis, this is likely to have led to random misdassification of exposure and thus a conservative bias, 36% of male referents had some asbestos exposure. However, in the area of Loudon where the study was made, exposure is known be high. Our findings must be considered with evidence from ocher epidemiological work. In a longftudinal study of male asbestos-cement workers in Louisiana, USA, Hughe* and Weill found no excess of lung cancer in workers without small opacities on the chest radiograph, even among long-term workers," although the power of the study was limited. In wprkers without small opacities followed for at least 20 years from first exposure, only 10 developed respiratory cancers, whereas 9*5 were expected. Siuis-Cremer and Bezuidenhout. in a necropsy series of 339 arrjphibole-asbestos miners in South Africa, found the proportional mortality ratio for lung cancer was raised in those with histological evidence of pulmonary fibrosis but not otherwise.'4 Evidence that fibrosis may not be necessary for asbestos-related lung cancer comes from a longitudinal study of Quebec asbestos miners and millers.The investigators concluded that most, but not all, lung Vol 345 -Apriltv, IMS JUL-13-2001 FRI 04:09 PM FAX MO, P, 18 THE LANCET cancers attributable to asbestos showed radiological abnormalities, most of which were small opacities. However, the chest radiographs used in this analysis were often taken many yean before death and some workers may subsequently have developed fibrosis. Our results are consistent with a hospital-based case-control study in which an excess risk of lung cancer was observed in patients with an occupational history of asbestos exposure but no pulmonary fibrosis.*1* M there is an association between asbestos and lung cancer without fibrosis, it is of interest for both scientific and practical reasons. From a scientific standpoint, it may suggest that asbestos Is not carcinogenic solely because it ft fibrogenic, although the absence of small opacities on the chest radiograph does not exclude flbrotlc change at a microscopic level. More importantly, ftom a practical point of view, it indicates that asbestos exposures which, do not cause small opacities on the chest radiograph may nonetheless increase the risk of lung cancer, and this need* to be considered by those responsible for industrial hygiene and compensation of workers with asbestosrelated iung disease. Ttw mtiywu lupportMl by a Ream ftom the Health and Safety toitCMlhra. References i MriDoirald JC, McDonald AO. Mesothelioma as an tacts* ofisfctsta* impest, tot Pew ft, Sdmeldsrinsa M, cdn. QinnriftoiMors of occupation*! ameer. Banbury report 9, Cold Spring Heritor, 1981; 75-W. Pi Vos Irvine H. Lament DW, Hole Df, Gliiis CR, Asbestos in lima cancer in the west of Scotland. BMJ J 993; 306:1503-06, 3 Doll R. Peto R. The catuo* of cancer. Oxford; Oxford Ucivairity Prate, 1981; 1238-45. 4 Office of Population Censuses and Surveys. Cancer srafisili*; rnistranone. England and Wales. Series MSI so 26. . 5 Meredith SfC, McDonald JC, Wortt-rdatcd respiratory Stoesit is the United Kingdom, 1989-1992; a report on the SWORD project, QtaMtA 1994(44: 183-89. 6 totcraationul Labour Office. Quiddinos for tha w* cf the 1LO initmidonal elawifieadon of radiographs of pnettmoeoajorst (sevised edition); occupational safety and health series no 2. Gmeva; International Labour Office, 1980. 7 McDonald AD, McDonald JC. Malignant mesothelioma is North America. Cimorr 1980; 46: 1650-54. 8 Liddell PDK. Radiological asiasement of pneum&conionc opaoties, SrJf tod Med 1977; 34: 85-94. 9 Mcleod TC, Carrington CB, Guentler EA, Diffuse MSersUrM disease: # new scheme ftr description, /ia&ototy 1983} 149:333-43. 10 KJptn HM, Lilia R, Smukl Y, Valciulw JA, Salitofni. Pulmonary fibrosis to asbestos insulation workers with iung cane#n a mdioiogica! and bistepathologital evaluation. BrJJntl Mxl 1983; 44t 96-jOO, 11 Bsctdak* MR. The epidemiology of wbeelos, to: lidded D, Mills, K, ed*. Mineral ffirare and heal*. Boca Raton: CRC Press, 1991:105, 12 Staples CA, Gamsu G, Ray CS, Webb WR, High maotoden computed tumography and lung function ffi asbestos-exposed w-ifkera with normal chest radiographs. Am ftwjw Dit 1989; 1.39* 1302-08. 13 Hughes JM, WcUl H. Asbyatosis as a precursor of wbestos-wtsted lung cancer rerulta ofa prospective mortality study, BrJjndMui 1991; 48: 29-33. 14 Sluia-Crtosor GK, BwuldeahoOtt BN. Rcbtior. t-eween asbestos and bronchial caaeor in amphibole asbestos mlato. MrJ todMed I98S: 4fe 537-49. 15 Liddell PDK, McDonald JC. Radiological findings a* predictor* of enortajiry in Quebec atbeato* wOrksra. BrJ hut Atof 1080* 37:257-61. 14 Mcrtischnlg M, Newell DJ, Barnsley WR, Cowan WK, Feinraann BL, Oliver E. Unsuspected oaposute to asbestos and breaedtegetsk o*tsr. SMJ1977; it 756-59. HIV I subtypes and male-to-femaie transmission in Thailand Chalyos Kunanusont, tgordis M Foy, Joan H Krei$s, Supachai Rerks-Ngarm, Praphan Ptianuphak, Smvanse Raktham, Chou-Pong Pau, Nancy L Young SomiMy we examined the risk faotore for heterosexual transmission of HIV In a case-control study of couples In Thailand. 90 HIV-positive men and their regular sex partners were enrolled at the Immune clinic, Chuiaiongkern Hospital, where 92% of male Index cases had HIV-I serotype A [subtype }. Most index cases had acquired HIV through sexual intercourse. 95 couples were enrolled at 15 detoxification clinics, where 7954 of them had HlV-l serotype 8 {subtype 8). Moat men had acquired HiV through injecting drug use (1DU), Afiw Bfefetoo, Department of Cwnmunleebfe Dboose Control, MMttiy of PubHe tfoetth, Neotheburi 11000, Thailand CC Kunanusont mb, S Rerks-Ngsrm Mbit University of Washington, -Seattle, USA [H M Foy kd, C Kunanuscnt, J K Krais* msth); Cfcdatontiwn Heepltail, Bangkok, Thailand (P pnanuphs* mo); Ban**** ANtrapoKten Administration (S Raktham wo); Centres tor &be<u Control end Prevention, Attonte, USA (C-PPauMp); eod theH(V/AlPSOotlelnrtlofl, Nontbaburi CHLYPungkiYCJiWSri . CofTc*pnr)dnc to; Or Chffiyo* Kunanusont ' iSSst-it..' .f'r The HIV seroconoordance rate was higher in the Immune clinic (6914) than in the IDU clinics {49% overall, and 21% after excluding female partners who were IDUs) (p*r0-01). The rate was also higher among couples in whom the male index case was infected with serotype A (subtype E) compared with serotype B (subtype B) [70% vs 52%, OR 2-1, 96% Ci 1-2-4-2). When we excluded couples in whom the female was also an IDU, the difference in concordance rates was even more pronounced (70% vs 26%, OR 6-8, 95% Cl 2-7-L7-6), Viral factors or subjects' characteristics may have contributed to the concordance rates, in a multivariate logistic regression analysis, HIV-i serotype A (subtype E) of male partners (adjusted OR 34, 95% Cl l-1-9-0) and history of IDU in female partners (adjusted OR 4-8, $5% Cl 1-4-15-9) remained independently associated with HIV seroconcordance. This study suggests that HIV-1 subtype i may be associated with higher risk of heterosexual transmission than subtype B. If so, the predominance of subtype E In Thailand may have contributed to the rapid spread of the HIV epidemto. Lsnoat 1995: 345:1078-83