Document MygKLv45k1MpxaeBKMxG49Ok

British Journal of Industrial Medicine 1983;40:340-345 Epidemiological survey of maintenance workers in London Transport Executive bus garages and Chiswick Works L RUSHTON, M R ALDERSON, AND C R NAGARAJAH From the Institute of Cancer Research, Sutton, Surrey SM2 5PX,UK ABSTRACT A mortality study of maintenance men employed for at least one year between 1 January 1967 and 31 December 1975 at 71 London Transport bus garages and Chiswick Works has been carried out. Over 97% of the population were successfully traced to determine their vital status at 31 December 1975.The mortality observed in the study population was compared with that which would be expected from the mortality rates for the all male population of England and Wales. The mortality of the study population from all causes was much lower than expected on this basis, as was the mortality from cerebrovasculardisease, ischaemicheart disease, and bronchitis. Mortality from all neoplasms was slightly less than expected overall and especially in the younger age groups. The observed deaths from cancer of the lung were approximately the same as those expected on the basis of national rates. Nevertheless, a deficit of observed deaths from lung cancer was obtained after adjusting for the higher mortality from this disease in Greater London. Raised mortality was found in subgroups of the population for several malignant disease groups but these were almost all based on small numbers of deaths, making it difficult to exclude chance effects.Both the number of men and deaths in the study were limited and the follow up time was also short. Considerable extension of the study to include more men and increase the follow up time would be required for any definite mortality patterns to emerge. ?pidemiological survey of ma jence among the diesel bus 8 higher than that expected am lation.' It was pointed out, ho buses had only been gradual1 1930s there may not have b of exposure or latency time tc the number of cases of lung Some preliminary results from a study of all cases of arising during service in five Transport workers, includinl ages and the central works. ed numbersof lung ca ed on the basis of 4 Le*deduced. The aim of this study was to examine the patterns of mortality by cause for maintenance workers at London Transport bus garages and Chiswick Works. This is the fourth part of a wider study of workers in the oil industry which has been carried out by the same research team.l-' As users of several of the products of the oil industry, it was suggested that the London Transport workers would provide a useful comparison group for the refinery and distribution centre workers. Further studies included a follow up study of those men who survived three months,' examination of sickness absence before the onset of coronary heart disease,' measurement of the physique of b u ~ m e n . 6a~nd the association of coronary artery disease with "stress."'1 A five year follow up ofa group of busmen with no evidence of ischaernic heart disease when first examined showed an assodation of ischaemic heart disease with high levels of either systolic blood pressure or plasma cholesterol.' Previous studies of London Transport workers There has been concern about the effect of diesel exhaust on health, particularly on the incidence of There has been a series of papers spanning several years reporting on work being camed out to investigate coronary heart disease in London Transport employees, particularly in bus drivers and conduc- tors.*-" In a study of workers during 1949 and 1950 it was first shown that coronary heart disease was less frequent in conductors than in bus driver^.^ lung cancer. Air pollution was measured in two London Transport diesel bus garages with only natural draft ventilation.j2 Concentrations of smoke were found to be higher inside the garages than Outside and high concentrations of smoke inside the garages correlated with periods of greatest bus movement; only small differences in the concentra- tions of hydrocarbon were found. Received 2 June 1982 The incidence of lung cancer among severa' categories of London Transport employees during Aaxpted 25 August 1982 1950-4 was examined and showed that the ina- 340 men surname, inil rr. .._ rorkers in and one year between 1 i and Chiswick Works Ed t o determine their ulation was compared I male population of was much lower than :haemic heart disease, overall and especially :re approximately the :it of observed deaths from this disease in on for several maligof deaths, making it ire study were limited t o include more men y patterns to emerge. a follow up study of e months,sexamination the onset of coronary it of the physique of ion of coronary artery ve year follow up of a evidence of ischaemic lined showed an associ:ase with high levels of or plasma ch~lesterol.~ wut the effect of diesel rly on the incidence of was measured in two ius garages with only incentrations of smoke e the garages than outs of smoke inside the n o d s of greatest bus mces in the concentrabund. :ancer among several mrt employees during showed that the inci- Epidemiological survey of maintenance workers in bus garages 34 1 dence among the diesel bus garage workers was no higher than that expected among the general popu- lation.' It was pointed out, however, that since diesel buses had only been gradually introduced since the 1930s there may not have been adequate duration of exposure or latency time to observe an increase in the National Health Service Central Register (NHSCR) at Southport. For all men found to be dead full details of death were obtained either directly from the NHSCR or from the Office of Population Censuses and Surveys. At all stages of the data preparation checks were the number of cases of lung cancer." carried out on the accuracy and plausibility of the Some preliminary results have been reported data and to eliminate duplicate records. from a study of all cases of lung cancer since 1950 The computer output for leavers between 1968 arising during service in five categories of London and 1975 also included men who had some form of Transport workers, including engineers in .bus garages and the central works.'. In each category the observed numbers of lung cancers were below those expected on the basis of Greater London death change of job status, including promotion and demotion, changing location within department, transferring to a different grade or department, and changing duties for medical reasons-that is, men rates. The highest mortality ratio was among bus who were not really "leavers." Nevertheless, many garage workers although this did not. differ significantly from the Underground motormen and guards who do not have any special diesel fume exposure. It was also pointed out that the smoking habits of the men were not known. Air pollution of them left London Transport subsequently. This resulted in duplicate entries being found for many individuals (about a quarter of all records on the computer output for leavers). It was assumed that most of the men would live in sampling was again carried out in the same two gar- the same area or close to their location of work. ages as before.'* The background concentration of Their area of residence was therefore coded to that benzo(a)pyrene in the outside air was found to be of the location of work using the OPCS area code." very much lower than in 1957 and the concentra- London Transport had used over 100 different job tions within the garages were also correspondingly titles in the study data and these were coded to 20 reduced. broader job groups (table 1). Supervisors and inspectors were combined for analysis, and there Method were also insufficient numbers of deaths for separate analyses for smiths, trimmers, carpenters, sheet The present study included all maintenance men metal workers, and workers in the experimental with year a length between of 1 of at least one january1967 and 31 c~on~tin~uo~usb~wseorvrkicseh,orpe.tiRreedas, oonr for leaving was coded to died other. The underlying cause in of 1975 at 71 London Transport bus garages and the death was coded using the eighth revision Of the engineering works at (-J,iswick. me bus garages are International Classification of Diseases.I6 collectively referred to as rolling stock by London The number Of years each man was in the study Transport and this term has also been used in this paper. Table 1 Distribution ofLondon Transportpopulation by For all men surname, initials, date of birth, date of job caregory (nor including unrraced men) joining, last or present job, and location of work were obtained from computer listings. For leavers, date of leaving and reason for leaving were also given. Alphabetical card index systems in the pension department were used to try and identify which of the retirees were still receiving a pension and which had died. For those who had died in service or after retirement and for men who had resigned, full name and last known address were obtained from an alphabetical card index in the personnel department. No National Insurance or National Health Service numbers were available. As it was thought that the information obtained from the pension department might be incomplete all leavers, except those who died in service, were Job carqory IS~u~~p~e~r~vi~rand or em^ Smth ~ ~ `kctrial ~ zs&Trimmer pinter No ~ ~uh Toolma& ~~~~~~~~d,~~dm~~~.fl RWCSSb n % (%) ~ : F r Itraced. Tracing was first carried out through the officeworker social security records at Newcastle. If this was hand unsuccessful a request for tracing was then made to ~ 342 Rushton, Alderson, and Nagarajah (person years) was calculated from his date of birth and dates of entry to and exit from the study using the man years computer language program.'' The total years were subdivided into five year age and calendar period groups. As all men in the study had a minimum of one year's service between 1 January 1967 and 31 December 1975 date of entry to the study was (a) 1 January 1968 for men who joined London Transport before 1967 and (b) one year after joining for those who joined after 1 January 1967. The expected number of deaths was calculated by applying the five year age and calendar period death rates of the comparison population to the person years. The comparison population used was the male population in England and Wales. Most men in London Transport live and work in Greater London but age and calendar period death rates by detailed cause were not readily available for this area. It was not thought appropriate to adjust the expected numbers of death using published standardised mortality ratios for Greater London as the numbers concerned were small. Differences in mortality between Greater London and the national population, however, have been taken into account when interpreting the results. The grouping of the causes of death was determined by that of the available comparison data. For 30 malignant causes the groups were those used in the serial mortality tables." For the non-malignant causes, data for the A list of 150 causes of death from the International Classification of Diseases were made available by the World Health Organisation. Colnparison of observed and expected deatbs The difference between the observed and expected deaths was examined by calculating a significance level (p value) for the observed numbers of deaths assuming it to be drawn from a Poisson distribution with mean equal to the externally calculated expected value. Two approaches to the analysis were used. Firstly, a search for pronounced excesses and deficits in numbers of observed deaths compared with those expected was made together with an examination of the data by location, age, and job. The significance levels obtained in these analyses were used informally as guidelines to indicate the magnitude and direction of the variations in mortality. The second approach included a priori testing of the findings from previous studies discussed in the introduction. Results Table 2 shows the status of the study population at the study end date, 31 December 1975. Only 2.2% of men were untraced. These have not been included in the analysis. There was a total of 50 008 person years of observation in the study with a mean follow up of 5.9 years. The total number of deaths was significantly lower than expected on the basis of the national po ulation (table 3). and especially for rolling stock PO = 495, E = 607.46, p < 0-OOOl).This deficit was also reflected overall in the numerically large nonmalignant disease groups including cerebrovascular disease, ischaemic heart disease, and bronchitis. Deficits were found in many subgroups of the population defined by location, age, and job for several of the non-malignant causes. The observed deaths from all neoplasms and cancer of the lung were approximately the same as those expected. There was only one excess, in cancer of the liver and gall bladder at Chiswick Works, although this was based on only four deaths (0= 4, E = 1-13, p < 0.05). Some excesses of observed deaths compared with those expected were found in particular job groups (table 4). Many of these are based on under five deaths and should thus be regarded cautiously. The wide confidence intervals shown in table 4 for some of these excesses reflects the uncertainty in the estimate of OIE which is based on such small numbers of deaths.'* Discussion Table 2 Status ofstudypopulation at study end date, 31 December 1975 IIn eAmepmlopromyemnetnt: Alive Ded. cause known Ded, cause unknown Emigrated Untraccd Total No 4671 3102 701 4 12 I94 8684 f %) (53.8) (35.7 8.1 (999 The mortality overall and from the numerically large non-malignant causes of death is much lower than that of the standard population with which it was compared. This is the type of result found in many studies that compare an industrially employed cohort with the national population which includes those in unstable employment or who are chronically ill or disabled. Previous studies have investigated coronary heart disease in London Transport workers. In the present demiologicd survey of n worker CW ;rlysiswere used. Firstly, xcesses and deficits in s compared with those with an examination of id job. The significance ilyses were used inforate the magnitude and I mortality The Second testing of the finding *ed in the introduction. the study population at rnber 1975. Only 2.2% 'hew have not been 'e was a total of 50 008 ~I the study with a mean , was significantly lower if the naoonal populafor rolling stock (0= 1). This deficit was also mencally large nonluding cerebrovascular &,ease.and bronchltls. ?ubgroupsof the popu', and job for several of in all neoplasms and le confidence i ~piciemiologicclslurvey of tnclititenatice workers in bus gclrages 343 j7)arbtnleai3nteOnabnsceervwedoraknedrseexmpeprlfoeydeddeaatthLso. OndIoEn, aTnrdapnsvpaolurtesbufosrgaalrlacgaeussaens,dalCl hnierowpilcuksmWsoarnkds causes from I with five or June 1967 mr to e deaths Lleceinber 1975 c -raiises (in ICD order) Observed deuhs All CaUSCS ~ nleolplasms Cancer of stomach Cancer of large intestine 705 216 26 IO Cancer of rectum 5 Cancer of liver and gall bladder 6 Cancer of ancreas 9 Cancer of rung and pleura I02 Cancer of prostate Cancer of urinary bladder II 12 cancer of brain and other parts of central nervous system 7 CLerurekbareomviaascular disease Chroruc rheumatic heart disease lschaenuc heart disease Other diseases of hean Hypertensive disease Diseases of artenes Other diseases of circulatory system pneumonia Bronchitis peptic ulcer Chronic and other nephritis Motor vehicle accidents 7 58 8 244 15 7 15 6 28 50 6 6 8 Expected deaths 836.56 226.77 25.5 I 13.19 9.92 4.09 9.75 101.31 8.93 8.64 5.77 4.65 74.4 I 8.74 276.63 20.33 13.06 17.05 7.87 32.01 6467 7.20 3.8 1 10.03 OIE 0.84 0.95 1.02 0.76 0.50 147 0.92 1-01 1.23 1-39 1.21 1.51 0.78 0.92 0.88 0.74 0.54 0.88 0.76 0.87 0.77 0.83 1.57 0.80 P <0oOOl 046 0.49 024 0 07 023 049 094 029 016 0.36 019 003 049 005 014 0.05 037 033 027 004 0.42 019 033 study there were significantly fewer deaths observed okerall than expected from ischaemic heart disease and especially in rolling stock and bus mechanics. Mortality from all neoplasms is only slightly less than expected overall. Nevertheless, a deficit was found for deaths under 55 years of age with the observed deaths approximately equalling those expected for deaths at older ages. An excess in welders and a slight excess in progress hands was found at Chiswick Works. Overall the observed deaths from lung cancer (102) are about the same as those expected (101) on the basis of national rates. As discussed above no adjustment was made for regional or social class variation in mortality. The SMR for lung cancer in Greater London, in which most of the London Transport men live and work, is 145-that is, the death rates from lung cancer are higher in Greater London than in the whole of England and Wales.*O A crude adjustment of the expected deaths calculated using national rates could be made as in the refinery study (adjusted expected deaths = expected deaths based on national rates x regional SMW 100). This would increase the overall expected number of lung cancer deaths to 117 giving a deficit (non-significant) of observed deaths. An excess was found in general hands in rolling stock and a regional adjustment would also reduce this excess. The mortality for these men would also not be greater than other men within their social class (V).z'There was also an excess from lung cancer in welders, although this was based on only three deaths. The Table 4 Observed dearhs (01.expected deaths ( E ) ,OIE. 95% confidence interval for OIE, andp values for excesses found in job groups L the numerically large 8i h is much lower than on with which it was result found in many . iidustnally employed dation Hhich includes t or who are chroni- c tgated coronary heart orkers In the present lntpector %elder Uelder Painter Painter Painter Progress hand Progress hand Bu\ mechanic O!hce worker Cffhce worker General hand General hand i I Rolling stock (RS)I Duearc group Chiswick W o r h ( C W ) O CW Cancer of liver and gall bladder 2 c w All neoplasms 5 c w Cancer of lung 3 RS Cancer of large intestine 2 RS Cancer of bladder 2 RS cw cw Pneumorua All neoplasms Cancer of bladder 3 I2 3 RS Cancer of brarn and central 4 nervous system CW All causes 16 CW cw Pneumonia Cancer of ancreas 2 3 RS Cancer of Pung 48 E 0.12 1.56 0.71 0.34 0.23 0.86 7-41 0.27 1.25 10.24 0.36 0.83 36.11 OIE 16.67 3-21 4.23 5.88 8-70 3.49 1.62 11.11 3.20 1.56 5.56 3.61 1.33 2.96 1.26 1.15 1-04 1.54 0.95 0.93 3.03 1.10 o.98 0 99 0.98 1.03 52,SY 6.74 10.93 18.55 27.44 9.02 2 63 28-71 7 32 2.37 17.54 9.33 1.69 <BO1 0.02 0.04 005 0-02 0.06 047 CO-01 0.04 0.06 0.05 0.05 0.03 344 above comments on regional variation in mortality apply to this excess. Other work has suggested at: association between welding fumes and impairec lung function,*zand lung cancer.?' A major problem in the intepretation of results for lung cancer is the lack of information on smoking habits. It is understood that at the end of the war about 15% of the bus fleet were still using petrol but these had all been replaced by diesel engines by 1950. Restriction on smoking was then lifted from 1950 onwards in all garages (PAB Raffle, personal communication). Several other excesses were found for various sites of malignant disease in subgroups of the population. Some of these, such as the excess from lung cancer in welders referred to above, have also been suggested in previous work-for example, cancer of the bladder in painters.z' As discussed earlier, however, they are all based on very small numbers of deaths, and the uncertainty in the estimate of the ratio of observed over expected deaths is reflected in the wide confidence intervals obtained for some of the results in table 4. Previous studies have shown an association between exposure to mineral oil and skin cancer. There was only one death from this disease, fewer than expected. The national age and calendar period mortality rates used for comparison reflect to some extent the changes in social habits and lifestyle over the eight years of the study. There is no evidence to suggest that these changes differ between the study and the comparison populations. The cause of death information obtained from the death details may be subject to inaccuracy through omission of information, misdiagnosis, or errors of classification and coding."Ls The difficulty of interpreting results from this type of study, where many comparisons of observed and expected deaths are made, is increased when, as in the present study, the numbers are small and the period of follow up is short. Altogether 71 bus garages were included, with the number of men eligible over the study period from the individual garages varying from six to 99. The numbers of men and numbers of deaths from each garage were thus rather small for formal analysis. Nevertheless, there was no indication from an informal examination of the deaths by bus garage that any individual garage was particularly concerned. Condusions This study of London Transport maintenance workers has not highlighted any problem a r e a s which might indicate that further investigation was needed. The information was limited, however, in both the Rushton, Alderson, and Nagarajah numbers of men in the study and the detail collected on each man. The duration of follow up was also, short, an average of only six years. The numbers of deaths by different causes were small, making it difficult to exclude chance effects. An extension of the duration of follow up and inclusion of leavers from additional years would increase the numbers of deaths, especially in causes with a long latent period, and might help to indicate whether any areas need further investigation. About half of the leaven in this study were men who left for reasons other than death or retirement and had to be traced through the national systems (a fairly lengthy procedure) to establish their vital status. Nevertheless, they contributed only 43 (6%) of the deaths. A useful study might thus be carried out more quickly using only records of men in service, retirees, and deaths in service and not tracing men leaving for other reasons if information was thought to be sufficiently accurate and complete. The data for this study were handled by Mrs Carol Fair and Ms Deborah Cummings. We benefited by help from many colleagues especially Dr Joan Davies and Mrs Doreen Folkes. The basic information was obtained from records in the London Transport administrative and pension departments and thanks go to all those personnel concerned. We are grateful to OPCS, DHSS, and the Registrar General for Scotland for tracing the leavers. The study benefited from the guidance of a working party of the Institute of Petroleum, in particular from Dr W L B Leese and Dr P A B Raffle. The main funding of the project was a grant from the Institute of Petroleum. The Cancer Research Campaign support of the division of epidemiology is also gratefully acknowledged. Requests for reprints to: Dr L Rushton, Thames Polytechnic, School of Mathematics, Statistics and Computing, Wellington Street, London SE18 6PF. References ' Rushton L. Alderson MR. An epidemiological survey of eight Oil refineries in Britain. Br I Ind Med 1981;38:225-34. 'Rushton L. Alderson MR. An cpidemidogicol survey of&/ dktibwion centra in Great Britain. Find report. London: Institute of Petrokum. 1982. ' Rushton L.Alderson MR. A case-control study to investigate the association between exposure to benzene and deaths frorn leukaemia in oil refinery workers. Br J Cancer 1981;43:77-84. M O ~ ISN, HCAY JA. ~ d PkAB. Coronary h e M disease and physical activity of work. k c r r 1953;i:1053-7, 1 1 1 1-20. ' Moms IN. Raffle PAB. Coronary heart disease in transPC"( workers. A progren report. Br J Ind Mrd 1954;11:26w- * MorrisJN.Heady JA. RIWt P a . PhysiqueofLondon busme"' Epidemiology of uniforms. Lance 1956;i:569-70. ' Morns JN, Kagan A, Pattison DC,Gardncr MJ. Raffle PAe' 5pidemiologica/ survey of n Incidence and prediction of isch busmen. L a n d 1966;i:5 53-9. ' Heady JA. Mom5 JN. Lloyd FJ. before the first clinical episode I Ind Mcd 1954;11:?04. 'Heady JA, Morris JN. Kagan A. disease in London burmen. A I reference IO physique. Br J Pic loRafne PAB. The health of 1 1 Y57 ;14:73-80. 1' Raffle PAB. Stress a5 a factor in C 11 Commins BT. Waller RE. Lawthe garages. Br J Ind Mcd 1957:14 "Schentcr MB. Dicsel exhaustIOM 1980;22:41-6. 'aWaller RE. Trends in lung can exposure to diesel fumes. t 1981;5:479-83. 15 offieof population Censuresand ing pkcc of residence andP ~ C Ct *World Health Or#aniution. clmsiFotion of diseases, i n W e 2. Geneva: WHO. 1967. II&non, and Nagarajah Epidemiological survev of maintenance workers in bus garages 345 Incidence and prediction of ischaemic heart disease in London busmen. Lancet 1966;1:55>9. " Heady JA. Moms JN, Lloyd FJ. Raffle PAB. Sickness absence before the first clinical episode of coronary heart disease. Er J Ind Med 1954;11:20-4. " Heady JA. Morns JN. Kagan A, Raffle PAB. Coronary heart disease in London busmen. A progress report with particular reference to physique. Br J Prev Soc Med 1961:15:14>53. "'Raffle PAB. The health of the worker. Er J Ind Med 1957;14:73-80. #'Hill ID. Computing man yean at risk. Er J Prev Soc Med 1973;26: 132-4. I' Case RAM, Coghill C, Davies JM, et al. Serial mortoliry tables, neoplastic discma. Vol 1. England and Wales I911-70. Lon- don: Institute of Cancer Research, 1976. le Bailar JC. Ederer F. Significance factors for the ratio of a Poisson , variable to its expectation. Biometrics 1964;20639-43. Office of Population Censuses and Surveys. Mortality sfarirtics area 1975. London: HMSO, 1977. (Series DHS No 2.) '' Office of Population Censuses and Surveys. Occupadonal mortal- han death or retir service and not tra I ' Raffle PAB. Stress as a factor in disease. Lancer 1959;i:839-43. I: Commins BT, Waller RE. Lawther PJ. Air pollution in diesel bus garages. Er J Ind Med 1957:14:232-9. Schenker MB. Diesel exhaust-an occupational carcinogen? JOM 1980;22:41-6. 1' Waller RE. Trends in lung cancer in London in relation to exposure to diesel fumes. Environmental International 1981;5:479-83. $9 Office of Population Censuses and Surveys. Area code for recordingplace ofresidence andplacc ofbinh. London: OPCS. 1974. 10 World Health Organisation. Manual of the international classification of diseasa. injuries and causes of deaths. Vols 1. 2. Geneva: WHO, 1967. ity 1970-72.London: HMSO. 1978. (Series DS No 1.) '* Arbarkhanzadeh F. Long-term effects of welding fumes upon respiratory symptoms and pulmonary function. JOM 1980;22:337-4 1. Beaumont JJ, Weiss NS. Lung cancer among welders. JOM 1981;23:639-44. Aldenon MR. The accuracy of the certification o f death and the classificnrion of the underlying cause of death from the death cenifiate. London: London University. 1965. (MD thesis.) I' Waldron HA. Vickentaff L. Intimations of quolity: ante-mortem and post-mortem diagnosis. London: Nuffield Provincial Hos- pitals Trust, 1977. e handled by Mrs mings. We benefit s especially Dr jrmiologicd survey of ei fed 1981;38:225-34. *&rn'dOgiCaI suw 'rirdn. Find report, rif'aDmrd study to tnvcstlgate to benzene and deaths fraj J. CRJ Cancer B Coronary heart disease 7 1953:rIOS>7, 111 ry heart disease m t J Ind Med 1954;ll:Z Phrpque of London c a 1956;1:569-70. C. Gardmr MI. Ra