Annals of Occupational Hygiene Advance Access originally published online on February 22, 2005
Annals of Occupational Hygiene 2005 49(5):367-373; doi:10.1093/annhyg/mei001
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
© 2005 British Occupational Hygiene Society Published by Oxford University Press
Original Article |
Mortality from Lung and Kidney Disease in a Cohort of North American Industrial Sand Workers: An Update
1 Department of Occupational & Environmental Medicine, National Heart & Lung Institute, Imperial College School of Medicine, Dovehouse Street, London SW3 6LY, UK; 2 Department of Biostatistics, School of Public Health and Tropical Medicine, Tulane University Medical Center, 1430 Tulane Avenue, New Orleans, LA 70112, USA; 3 Department of Environmental Health Science, School of Public Health and Tropical Medicine, Tulane University Medical Center, 1430 Tulane Avenue, New Orleans, LA 70112, USA; 4 Department of Medicine, School of Medicine, Tulane University Medical Center, 1430 Tulane Avenue, New Orleans, LA 70112, USA
* Author to whom correspondence should be addressed. Tel: +44 (0)20 7351 8934; fax: +44 (0)20 7351 8091; e-mail: c.mcdonald{at}imperial.ac.uk
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Background: A previously published cohort study of some 2670 employees of the North American sand industry, followed through 1994, provided strong evidence of a causal relationship between quartz exposure and death from both silicosis and lung cancer, after allowance for cigarette smoking and in the absence of known occupational carcinogens. Unexpectedly, a significant excess mortality from chronic non-malignant renal disease [observed 16; expected 7.6; standardized mortality ratio (SMR) 212] was also found, whereas deaths from renal cancer at this stage were close to expectation (observed 6; expected 5.2).
Objectives: Our primary aim was to discover whether death from chronic renal disease was related to the estimated intensity of crystalline silica exposure. A further aim was to determine whether or not our previous estimates of lung cancer and silicosis risk were confirmed by mortality in the cohort 6 years later.
Methods: With help from the US National Death Index, surviving members of the cohort, with the exception of employees of a small plant in Canada, were traced through 2000. The cause of death was determined for all who had died, for comparison against National and State mortality rates. Nested case-referent analyses were then undertaken, as previously, of deaths from lung cancer and silicosis, plus end-stage renal disease and kidney cancer, in relation to quantitative re-estimates of quartz exposure.
Results: The total number of deaths through 1994 was 990; there were 231 additional deaths during the period 1995–2000. The SMRs were significantly higher in the later than the earlier period, mainly due to a relative increase in heart disease and external causes. The updated odds ratios for lung cancer and silicosis were almost identical to those published previously, with lung cancer risk again related to average silica concentration and cumulative exposure, but not to length of employment. In contrast, risks of neither end-stage renal disease nor renal cancer were related to cumulative exposure, although now based on 19 cases (SMR 239), and 10 cases (SMR 202), respectively, in fact, opposite trends were apparent for both diseases. However, because of the small numbers there was only limited power to assess the statistical significance of these trends or of any separate relationship with the duration or intensity of exposure.
Conclusions: Our findings support a causal relationship between lung cancer and quartz exposure after allowance for cigarette smoking, in the absence of other known carcinogens, but failed to find similar evidence to explain the excess mortality from either chronic renal disease or kidney cancer.
Keywords: lung cancer • renal disease • sand • silica • silicosis
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Findings from a cohort of 2670 men employed in nine North American sand-producing plants and followed through 1994 showed significant excess mortality from lung cancer (McDonald et al., 2001
). After allowance for a strong effect of cigarette smoking, odds ratios for lung cancer in a nested case-referent analysis were related to cumulative crystalline silica exposure and average concentration, but not to length of employment (Hughes et al., 2001; Rando et al., 2001). It was concluded that these findings supported a causal relationship between lung cancer and quartz exposure in the absence of other known carcinogens. Somewhat unexpectedly, however, there was a strong indication in this cohort of excess mortality from chronic non-malignant renal disease [16 deaths; standardized mortality ratio (SMR) 212; P = 0.002], thought worthy of further investigation. Therefore, mainly for this purpose survivors from this cohort were subject to a further 6 year follow-up of mortality until the end of 2000. |
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Cohort mortality
The cohort initially comprised 2670 men ever employed for at least 3 years, with 1 month or more during the period 1940–79, in nine selected North American sand-producing plants. As one plant, with 215 men, was in Canada, for simplicity further study was confined to the 2455 men from the eight US plants. However, two employees of the latter were identified as duplicates and one other had terminated employment in 1936, leaving a final total of 2452 employees for the update.
With help from the US National Death Index, vital status was determined until the end of 2000 in all but 21 men (0.9%): 1205 (49.1%) as alive and 1225 (50.0%) as deceased. The cause of death, coded according to the International Classification of Disease (ICD) in effect at the time, was obtained for all men who had died. Observed deaths by cause were compared with numbers expected from US mortality rates and, for selected causes of death, against State rates.
Case–control analyses
Cases were defined for this purpose as all deaths from lung cancer, silicosis (or silicotuberculosis), kidney cancer and chronic renal disease (nephritis or nephrosis). For each case individually, two controls (referents) were sought from cohort members employed at the same plant, born within 5 years (three if possible), first hired within 5 years (three if possible) of the case and who survived the case. At least one control was identified for 108 of the 112 lung cancers, 32 of the 37 silicosis/silicotuberculosis cases, all 19 of the chronic renal cases and all 10 kidney cancer cases. As in the previous study, job and smoking histories were obtained from employment records for the newly identified cases and controls of lung cancer and silicosis, and job histories only for renal disease, all without knowledge of the case/control status. Overall, the number of cases for which job histories were obtained for both the case and at least one corresponding control were 105 for lung cancer, 29 for silicosis, 13 for chronic renal disease and 9 for kidney cancer.
Exposure estimation
Exposure levels from 1975 to the present were derived for each plant from existing company-owned records of respirable silica exposure and respirator use. The arithmetic mean exposure level for each job for each year was determined and then adjusted for the observed frequency of use of respiratory protective equipment for that job in that year, assuming a protective factor of 5 (80% reduction in exposure) during use. The protective factor of 5 was derived from a survey of published data on workplace protection factors for air purifying respirators used for protection against industrial dusts (Rando et al., 2001). Significant use of such protection began during 1978 and steadily increased in frequency thereafter, with certain jobs reporting continuous use in the latter period of the study. For periods before 1975, current exposures, unadjusted for respiratory protection, were extrapolated back to the date of any relevant change in process, work practice or engineering control. For years prior to any change, exposure estimates were based on particle counts collected by Hatch in the industrial sand industry, 1946–1955, converted to gravimetric concentrations of respirable silica by an algorithm derived specifically for this industry (Rando et al., 2001).
The average concentration of respirable crystalline silica by year for jobs appearing in the work histories of cases and controls is shown in Fig. 1. Exposure levels in the early years were generally high, often surpassing 400 µg m–3. In the early years of the study, there were few subjects, most of whom happened to be in jobs with low estimated exposure. Beginning around 1945, a steady decline was observed as improvements in engineering controls were instituted at the plants. Between 1977 and 1979, the average exposure levels dropped abruptly from 
150 to <100 µg m–3, mainly as a result of increased use of respiratory protective equipment. The apparent increase from 1996 onwards was derived from a very few subjects, who by chance were in jobs with higher than average estimated exposure.
|
Statistical methods
Members of the cohort were followed, as before, from 3 years after the date of hire or from 1 January 1940, and person-years were calculated until death, age 95 or 31 December 2000 or, if unknown, from the date last known to be alive. One-tailed tests were used for assessing the statistical significance of elevated SMRs based on observed numbers, following a Poisson distribution, and expected numbers from either US or State rates. Poisson regression was used to test differences or trends in SMRs across categories. Differences in smoking and exposure between cases and controls were analysed by conditional multiple logistic regression, using the phreg procedure of SAS. The same boundaries were used, as before, to define exposure categories. Cigarette smoking was taken into account by including three dummy variables in the model (ever smoker, never smoker or unknown).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Overall mortality
SMRs 20 or more years after hire were compared for selected causes for the earlier follow-up period (through 1994) and for 1995–2000 (Table 1). The SMR for all causes combined was significantly higher for the later period than previously (142 and 118, respectively). For individual causes of death, the SMRs for all causes and for external causes were significantly higher during the later period than previously, while that for heart disease was marginally so. For all other causes, differences by follow-up period did not approach statistical significance (all P-values >0.20). While SMRs using State rates were generally lower than those using US rates, the same causes of death were significantly elevated. For example, SMRs 20 or more years from hire, based on US and State rates, respectively, were 147 and 137 for lung cancer, 164 and 151 for non-malignant respiratory disease (NMRD), 202 and 197 for kidney cancer and 280 and 222 for nephritis/nephrosis.
|
When examined by decade in which death occurred (not shown), SMRs for lung cancer have been relatively stable since 1970, those for NMRD were highest before 1970, with silicosis risk being highest before 1970 and relatively stable since. By year of hire (results not shown), nephritis and renal cancer SMRs followed a generally decreasing trend over time; the highest risks from these two diseases were in men hired before 1950. Lung cancer SMRs were elevated in all plants, except two in Ohio where risks due to both silicosis and NMRD were higher than elsewhere. Excess risk from nephritis/nephrosis was observed in most plants, but from renal cancer mainly in Pennsylvania and West Virginia.
Case–control analyses
Silicosis and lung cancer
As the number, but not identity, of deaths available for the present analysis was the same for silicosis as in the previous study and, for lung cancer, increased only from 91 to 108, findings on exposure–response were likely to be similar, though not necessarily identical. Nevertheless, the extent to which the update has strengthened or weakened the earlier conclusions on these two diseases remains important. A comparison of the main findings on silicosis mortality originally and now is set out in Table 2. The odds ratios shown were adjusted for matching and smoking category and were relative to the lowest exposure category. Taking into account the very small numbers, the original and updated ratios agree reasonably well, especially in those lagged 15 years, which demonstrated a clear and statistically significant trend. Further analysis of more recent data showed that risk of death from silicosis was independently related to both duration and intensity of exposure, and was strongly related to smoking (odds ratio for smokers relative to non-smokers 5.09; 95% CI 1.29, 20.07).
|
A comparable examination of similarly adjusted lung cancer risks in relation to cumulative exposure at the two phases of the study is presented in Table 3. Once again the two sets of ratios are very similar, with little to choose between those with lagged or unlagged exposure; in both, the trend was statistically significant. As had been observed previously, further analysis of data from the updated study indicated that the risk of lung cancer was related to average silica concentration, but not to length of employment, and strongly and independently related to smoking category (most recent odds ratio for smokers versus non-smokers was 5.37; 95% CI 2.33, 12.35). There was a significant trend (one tailed P = 0.015) in risk across silica concentration categories (see Table 4) in which the mean values for years employed were very similar, whereas risk was unrelated to years employed after adjusting for exposure concentration (see Table 5).
|
|
|
Renal disease
The case-referent analysis of chronic renal disease and kidney cancer were based on all 19 cases of the former, with 34 matched controls, and on all 10 cases of the latter, with 20 matched controls. However, although similar in terms of years employed, work histories were obtained for only some 60% of the chronic renal disease case and control pairs, but 90% of the kidney cancer cases. Tables 6 and 7 present the results of the analysis of these two diseases in relation to cumulative exposure. Whether lagged or unlagged, the findings in both were very similar, with decreasing trends in odds ratio across the four categories of cumulative exposure; these trends were not statistically significant. Very small numbers did not allow a more detailed analysis to assess the separate contribution to risk of duration and intensity of exposure (details available on request).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
This study of North American industrial sand workers was initiated in 1990, because of the dearth of epidemiological evidence then available as to whether exposure to crystalline silica was associated with increased risk of lung cancer, after allowance for cigarette smoking and in the absence of other known occupational carcinogens. The conclusion reached, based on mortality through 1994 and confirmed by the update through 2000, was that after accounting for a strong effect of cigarette smoking and infrequent asbestos exposure, lung cancer risk was significantly related to average concentration of crystalline silica and to cumulative exposure, though not to length of employment. Our information on smoking was admittedly incomplete and less than perfect. However, while the data were incomplete, the results for lung cancer (a significantly elevated odds ratio of 5.4 for ever smokers versus never smokers) support the validity of the data available regarding ever/never smoking.
Silicosis mortality was, as expected, strongly exposure-related; the present study, with mortality carried through to the end of 2000, has replicated these earlier findings. In the absence of radiographic data, the extent to which the excess risk of lung cancer was confined to men with pulmonary fibrosis could not be assessed. The main purpose of the update, however, was to examine the earlier finding of significant excess mortality from chronic non-malignant renal disease, based now on 19 deaths (SMR 239), and a small but definite excess of renal cancer, now based on 10 deaths (SMR 202), while adding to the information on exposure–response for lung cancer and silicosis in this cohort.
It is evident from case–control analysis of the two renal diseases that any relationship to employment in the industrial sand industry was not apparently related to either cumulative exposure or average intensity in the same way as silicosis or lung cancer. On the other hand, the risk of diseases attributable to biopersistent agents, or those which entail immunologic sensitization or analogous pathophysiological change, may well present a different pattern. Our data on end-stage renal disease are in line, nevertheless, with some but not all findings of Steenland et al. (2001, 2002), but not with studies of Californian diatomite workers (H. Checkoway, personal communication), Vermont granite workers (Attfield and Costello, 2004) or, so far as we are aware, any of the other cohorts cited by the IARC in 1997, including British pottery workers (N. M. Cherry, personal communication). However, a variety of so-called autoimmune disorders, including chronic renal disease, have been frequently attributed to silica exposure (Steenland and Goldsmith, 1995; Steenland et al., 2001). For example, an exceptionally high prevalence of non-organic-specific auto-antibodies has been found in silicotic sandblasters, including antinuclear antibodies (Jones et al., 1976). An immunological mechanism is not incompatible with our own findings.
The small and unforeseen excess of renal cancer which has now appeared is more difficult to explain, unless aetiologically related in some way to chronic renal disease. With the possible exception of a small excess reported by Attfield and Costello (2004) in Vermont granite workers (observed 18; SMR 137), renal cancer has not been related to silica exposure previously. Our finding therefore remains unconvincing, although sufficient to deserve testing against a prior hypothesis in future studies.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Our findings again support a causal relationship between lung cancer and quartz exposure after allowance for cigarette smoking, in the absence of other known carcinogens, but failed to find similar evidence to explain the excess mortality from either chronic renal disease or kidney cancer.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
This study was supported by a grant from the National Industrial Sand Association and the National Mining Association. The authors are grateful to personnel from participating companies for access to work and medical records and exposure data.
Received October 6, 2004; in final form January 13, 2005
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Attfield MD, Costello J. (2004) Quantitative exposure-response for silica dust and lung cancer in Vermont granite workers. Am J Ind Med; 45: 129–38.[CrossRef][Web of Science][Medline]
Hughes JM, Weill H, Rando R et al. (2001) Cohort mortality study of North American industrial sand workers. II. Case-referent analysis of lung cancer and silicosis deaths. Ann Occup Hyg; 45(3): 201–7.
IARC. (1997) Silica, some silicates, coal dust and para-aramid fibrils. IARC monographs on the evaluation of the carcinogenic risks to humans, Vol 68. Lyon: International Agency for Research on Cancer.
Jones RN, Turner-Warwick M, Ziskind M et al. (1976) High prevalence of anti nuclear antibodies in sandblasters' silicosis. Am Rev Respir Dis; 113: 393–5.[Medline]
McDonald AD, McDonald JC, Rando R et al. (2001) Cohort mortality study of North American industrial sand workers. I. Mortality from lung cancer, silicosis and other causes. Ann Occup Hyg; 45(3): 193–9.
Rando R, Hughes JM, Weill H et al. (2001) Cohort mortality study of North American industrial sand workers. III. Estimation of past and present exposures to respirable crystalline silica. Ann Occup Hyg; 45(3): 209–16.
Steenland K, Goldsmith DF. (1995) Silica exposure and autoimmune diseases. Am J Ind Med; 28: 603–8.[Web of Science][Medline]
Steenland K, Sanderson W, Calvert GM. (2001) Kidney disease and arthritis in a cohort study of workers exposed to silica. Epidemiology; 12: 405–12.[CrossRef][Medline]
Steenland K, Rosenman K, Socie E et al. (2002) Silicosis and end-stage renal disease. Scand J Work Environ Health; 28(6): 439–42.[Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  B Vacquier, S Caer, A Rogel, M Feurprier, M Tirmarche, C Luccioni, B Quesne, A Acker, and D Laurier Mortality risk in the French cohort of uranium miners: extended follow-up 1946-1999 Occup. Environ. Med., September 1, 2008; 65(9): 597 - 604. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. Yu, L. Tse, C. Leung, T. Wong, C. Tam, and A. Chan Lung cancer mortality among silicotic workers in Hong Kong--no evidence for a link Ann. Onc., June 1, 2007; 18(6): 1056 - 1063. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L A Tse, I T S Yu, C C Leung, W Tam, and T W Wong Mortality from non-malignant respiratory diseases among people with silicosis in Hong Kong: exposure-response analyses for exposure to silica dust Occup. Environ. Med., February 1, 2007; 64(2): 87 - 92. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Pelucchi, E. Pira, G. Piolatto, M. Coggiola, P. Carta, and C. La Vecchia Occupational silica exposure and lung cancer risk: a review of epidemiological studies 1996-2005 Ann. Onc., July 1, 2006; 17(7): 1039 - 1050. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||