Registration Dossier

Administrative data

Description of key information

Human epidemiological studies indicate a causal relationship between benzene exposure and acute non-lymphatic leukaemia.  

Additional information

Repeated Dose - Haematotoxicity

It has been clearly established for many years that chronic benzene exposure leads to depression of red and white blood cells (EU RAR, 2008). On the basis of a cross-sectional study by Rothman et al. (1996) which reported significantly reduced lymphocyte counts for a group of 22 workers exposed to benzene concentrations between 1.6 and 30.6 ppm compared to a group of 44 non-exposed workers, the EU RAR (2008) concluded that the most sensitive reaction in humans to chronic benzene exposure is lymphopenia and that an overall LOAEC for blood cell depression is suggested to be 32 mg/m³ (10 ppm). The EU RAR (2008) derived a NOAEC of 3.2 mg/m3(1 ppm) for depression of lymphocytes by benzene taking into consideration information on changes in lymphocyte counts from all the available studies. Lymphopenia was also considered by the US EPA (2002) to be the most sensitive toxic endpoint for humans exposed to benzene and Rothman et al. (1996) was selected as the critical study to derive an inhalation reference concentration. Benchmark dose modelling gave a 95% lower confidence unit for the benchmark concentration of 7.2 ppm, which is considered by US EPA to be an adverse-effect level. However, since this study is based on very few workers, there is considerable uncertainty regarding the concentration of benzene that is responsible for haematotoxic effects in humans.

Since the EU and EPA risk assessments were performed, additional studies have been completed which give better information about the response curve at low concentrations. The key studies are studies by Qu et al. (2002, 2003), Schnatter et al. (2010), Swaen et al. (2010) and Koh et al. (2015). The key study byQu et al. (2002, 2003) compared haematological parameters in a group of 130 exposed workers with median benzene exposure of 3.8 ppm in the four weeks before their haematological test with those among 51 unexposed workers. There were significant exposure-dependent decreases in RBCs, WBCs, and neutrophils, but there was no clear benzene effect on lymphocytes or platelets. The authors reported significantly reduced total WBCs, neutrophils, and RBCs in a small subgroup of 16 female subjects exposed to < 1 ppm. However, the reductions in the counts of total WBCs, neutrophils, and RBCs within this subgroup were much greater in magnitude than those reported for the main (>0–5 ppm) exposure group, rendering the results in this subgroup analysis of questionable value. An LOAEC of 2.27 ppm (mean exposure during the 4 week period preceding and including the day that blood sample taken) is obtained based on effects seen for RBC, WBC and neutrophils in the < 5 ppm exposure groups.

 

Another key study is by Schnatter et al. (2010) who examined haematological parameters from 928 workers exposed across a wide range of benzene concentrations (median 7.4 mg/m3 (2.3 ppm); interquartile range, 0.9 mg/m3 (0.3 ppm) - 9.2 ppm (29.5mg/m3)) and 73 unexposed workers. Benzene exposure was assessed via more than 2900 individual monitoring readings. Change point regressions were fitted to the data which indicated that the most sensitive parameters to benzene appeared to be neutrophils and the mean platelet volume (MPV), where change points were estimated to occur at benzene air concentrations of 7.8 – 8.2 ppm. However, there is no support in the literature for the finding that MPV is a particularly sensitive endpoint. There is some support in the literature for an effect on neutrophils, and Qu et al. (2002, 2003) reported a 12% reduction in neutrophil counts among workers exposed to 0-5 ppm benzene in the previous four weeks, although higher past exposures may have influenced these findings. Other analyses performed to assess dose response for the out-of-range indices showed a significantly increased risk for the exposure categories of <1 ppm and >10 ppm, but the dose-response was irregular. Hence the one-sided 95% lower confidence limit for the neutrophil change point (3.5 ppm) derived from the study by Schnatter et al. (2010) can be regarded as the NOAEC in the present study.

Another key study is by Swaen et al. (2010) in which the authors analysed haematological data from an exposed group of 701 workers (8,532 blood samples) and a non-exposed group of 1,059 workers (12,173 blood samples). The study extends an earlier investigation by Collins et al. (1997) which only considered the prevalence of abnormal haematological results. The study by Swaen et al. (2010) is much larger and has much greater power to detect effects because it analyses continuous endpoints. The blood sample of an individual was linked with the mean exposure of individuals performing the same job at the time the sample was taken. The average exposure estimate for all blood samples in the exposed group was 0.22 ppm, ranging from 0.01 to 1.85 ppm. A stratification of the exposed population into three subgroups (<0.5 ppm, 0.5-1 ppm and >1 ppm) showed significantly reduced levels of eosinophils for exposure levels above 0.5 ppm, but no difference for any of the other haematological parameters between the three exposure categories or when compared with the non-exposed group. However, Schnatter et al. (2010) reported no association between eosinophil counts and benzene exposure and Qu et al. (2002, 2003) provide no support for an effect below 15 ppm. The absence of haematological effects at low exposures is supported by another large study by Tsai et al. (2004) who found no differences after comparison of haematological surveillance results for 1200 workers with low benzene exposure (0.6 ppm 8 hr TWA between 1977 and 1988; 0.14 ppm 8 hr TWA since 1988) with those of 3,227 non-exposed employees.

 

Another key study by Koh et al (2015) examined the relationship between low-level benzene exposure and blood cell counts and included 21,140 blood samples from 10,702 Korean workers whose first compulsory health examination for benzene exposure occurred between 2005 and 2008. However, lack of exposure before 2000 could not be confirmed. Blood cell counts of benzene-exposed workers were extracted from a nationwide database containing records of health examinations. Mean benzene exposure was estimated for workers with various combinations of factory/industry/process codes at the time of examination using means calculated using 8,679 personal air benzene measurements recorded between 2004 and 2008. Benzene exposure of blood samples was categorised as <0.01 ppm (27.2%);≥0.01 to <0.1 ppm (36.4%);≥0.1 to <0.5 ppm (27.3%); and≥0.05 (9.1%). There was some evidence of an effect of exposure on RBC counts of male workers, but the findings from analyses of categorical and continuous measures of outcome were not consistent. Mean RBC counts of male workers were significantly reduced in exposure categories of≥0.01 to <0.1 ppm and≥0.1 to <0.5 ppm, but not≥0.05 ppm, whereas Odds Ratios (OR) of an abnormally low RBC count in male workers were significantly increased in the≥0.05 ppm category, but significantly reduced in the exposure category of≥0.01 to <0.1 ppm. Spline regression analyses were also not consistent with categorical analyses based on abnormal results, and RBC counts in female workers did not show similar effects. There was little evidence of adverse effects of benzene exposure for other parameters.

A study by Lan et al. (2004) although used by the US ATSDR (2007) to derive a chronic-duration inhalation Minimum Risk Level (MRL) for benzene, is considered to be of lower reliability compared to the key studies by Qu et al (2002, 2003), Schnatter et al. and Swaen et al. The study compared 250 subjects from two shoe factories with140 unexposed controls. Subjects were categorised into four groups by mean benzene levels measured during the month before phlebotomy (control; <1 ppm; 1 to <10 ppm; and≥10 ppm). The authors reported significant depressions in WBC, granulocytes, lymphocytes, CD4 T cells, B cells, monocytes, and platelets at benzene concentrations less than 1 ppm based on a group of 109 subjects categorised as exposed to < 1 ppm (mean exposure 0.57 ppm). However, it is clear from an accompanying report on exposure in the different workgroups (Vermeulen et al., 2004) that most of the 109 subjects would have been categorised as higher exposed had the investigators used a job exposure matrix to assign workers to groups. Furthermore, over half of the workers categorised as exposed to < 1 ppm had a past cumulative exposure of over 40 ppm-years, although the average length of employment was only 6.2 years, and many workers were also exposed to a range of other chemicals, including toluene, pentane, ethyl benzene, xylene, trichloroethane and heptane. A small group of 30 cutting workers was identified whose mean exposure to benzene was <1 ppm and with negligible exposure to other solvents. Decreased levels of WBCs, granulocytes, lymphocytes, and B cells compared to controls (P < 0.05) were reported. However, only limited weight can be given to this finding given the contradictory evidence from the very much larger studies by Swaen et al. (2010) and Tsai et al. (2004) of groups of workers exposed to similar levels of benzene at the time of their haematological test, and the study by Pesatori et al. (2009). Indeed, Koh et al. (2015) in analyses that included all subjects, reported significantly increased levels of WBC and lymphocyte counts in all exposure categories and significantly elevated levels of neutrophils in the≥0.05 ppm category.

 

The key studies are those by Qu et al (2002, 2003), Schnatter et al. (2010), Swaen et al. (2010), and Koh et al. (2015). These studies provide no evidence of effects below 2.27 ppm, and studies of lower reliability also do not support the finding by Lan et al. (2004) of effects at exposure levels lower than this. A LOAEC of 2.27 ppm (7.3 mg/m3) is obtained based on effects for RBC, WBC and neutrophils reported by Qu et al. (2002, 2003).

Carcinogenicity

It has been clearly established for many years that exposure to benzene causes acute myelogenous (non-lymphocytic) leukaemia (AML or ANLL) and a variety of other blood-related disorders in humans (IARC, 1982). Some epidemiology studies have suggested that exposure to benzene may also cause other forms of haematopoietic cancer; however IARC (Baan et al, 2009) has recently concluded that, although there is sufficient evidence for an increased risk of AML/ANLL in humans, there is onlylimitedevidence of carcinogenicity in humans for acute lymphocytic leukaemia (ALL), chronic lymphocytic leukaemia (CLL), multiple myeloma and non-Hodgkins lymphoma (NHL) andinadequateevidence for chronic myeloid leukaemia (CML). There is also sparse literature on specific myeloid tumours, such as myeloproliferative disorders (MPD) and myelodysplastic syndromes (MDS), which can precede and evolve under certain conditions into AML. IARC (2010) did not mention these myeloid tumours in their recent evaluation of benzene carcinogenicity, but did state thatbenzene is carcinogenic to the bone marrow causing leukaemia and myelodysplastic syndromeswhen discussing the leukaemagenic potential of benzene. Recently an association between MDS and benzene exposure has been reported by Schnatter et al. (2012) in a study of petroleum distribution workers exposed to relatively low levels of benzene, Schnatter et al (2012) reported associations between MDS and cumulative benzene exposure and peak exposure to benzene (at least 1-year employment in jobs likely experiencing >3 ppm exposure for 1560 minutes at least). The latter association appeared to be more robust, but it was noted that it is difficult to ascribe precise concentrations of benzene to MDS. A recent study by Copley et al. (2017) has confirmed the association between MDS and benzene exposure, especially for certain subtypes. In analyses that included all MDS subtypes, significantly elevated associations were reported for exposures > 3 ppm, although the report of the exposure assessment noted that exposure levels were not numerically precise.

 

A number of epidemiology studies provide clear and consistent evidence of a causal association between benzene exposure and AML or ANLL (Schnatter et al. 2005), but few provide information which is useful for quantitative risk assessment (Vlaanderen et al. 2008). The cohort used most widely for the purpose of risk assessment is the study of Pliofilm workers at three facilities (Rinsky et al. 1981, 1987, 2002; Paxton et al. 1994a). The key risk assessment by Crump (1994) was based on the update by Paxton et al. (1994a) which included 1212 workers followed from 1940 through 1987. The study has adequate power, long follow up and few reported co-exposures to carcinogenic substances relative to other studies. However, there has been considerable debate about the exposures of these workers. Estimates of exposure to benzene of workers in the cohort were derived by Rinsky et al. (1987), but other estimates were developed by Crump and Allen (1984) and Paustenbach et al. (1992) who both concluded that Rinsky et al. (1987) had substantially underestimated the benzene exposure. This would likely have resulted in an overestimation of the risk of leukaemia at low levels of exposure. However, Utterback and Rinsky (1995) dispute that Rinsky et al. (1987) had underestimated exposure. A further set of estimates developed by Willliams et al.and Paustenbach (2003) has not yet been used to estimate excess risk, but has been used by Rhomberg et al (2016) to calculate SMRs for acute nonlymphocytic leukaemia (ANLL) in the Pliofilm cohort, using lifetable analyses. Williams and Paustenbach (2003) suggest that Rinsky et al. (1981, 1987) under-predicted benzene exposures for most jobs, Crump and Allen (1984) both under- and overpredicted benzene exposures, and Paustenbach et al. (1992) generally over-estimated exposures for those job categories that had the highest exposure by about a factor of two to four.

 

While there have been many attempts made to derive risk estimates using information from the Pliofilm study, the study by Crump (1994) forms the basis of many regulatory assessments including WHO (2000) and US EPA (1998). Crump (1994) presented 24 dose-response analyses which modelled the association between all leukaemias and acute myelocytic or monocytic leukaemia combined (AMML) and benzene exposure. The models assessed the impact of factors such as (i) additive or multiplicative response, (ii) six types of functional form: a linear or non-linear function of either cumulative exposure or weighted cumulative exposure, and an intensity dependent model which accumulates an unweighted or weighted non-linear function of intensity over the duration of exposure, (iii) two sets of exposure measurements (Crump and Allen [1984] vs. exposure estimates by Paustenbach [1992]). These model parameter estimates remain relevant as the latest study update by Rinsky et al. (2002) indicated that only one additional new death from leukaemia had occurred in the cohort. However, the estimates of added deaths from leukaemia or AMML derived by Crump (1994) for occupational exposure of 1 ppm (45 years) range over more than two orders of magnitude (0.02 to 5.1 additional leukaemia deaths attributable to benzene per 1000 individuals with 45 ppm-years of cumulative benzene exposure). A mid-range position for leukaemia is suggested by Paxton et al. (1994b) who derived estimates of excess deaths using Cox proportional hazards models, an approach which has since become the preferred approach of many epidemiologists to model survival data. Using the same exposure estimates as Crump (1994), they derived estimates of excess leukaemia deaths (0.3 to 0.5 additional leukaemia deaths attributable to benzene per 1000 individuals with 45 ppm years of cumulative benzene exposure). TCEQ (2007) has recently calculated extra risk estimates for AMML and leukaemia for two of the models derived by Crump (1994). They assumed 70 years exposure and used much more recent US mortality and survival rates than Crump (1994), and obtained estimates of extra risk that were more than two fold lower than the added risk estimates of Crump (1994).

 

The wide variation in excess risk estimates is probably due to the fact that there is limited information to distinguish between the different modelling approaches, and potential thresholds for effect. The recent analysis by Rhomberg et al (2016) which is based on the Williams and Paustenbach (2003) exposure estimates, strongly suggests that effects only occur at levels of exposure > 80 ppm-years. The SMR analysis for ANLL by quintiles of cumulative exposure based on Williams and Paustenbach estimates showed a large excess (SMR = 9.94; 95% CI 4.0020.48) for the > 80.11 ppm-years category. There was only 1 ANLL death with lower cumulative exposure (63.7 ppm-years accumulated over only 1.5 years of employment) and 2.1 expected deaths. Furthermore, the average exposure of all ANLL cases exceeded 14.9 ppm for Williams and Paustenbach estimates, although the expected number of deaths was not reported for workers exposed to < 14.9 ppm. The exposure of many workers was low as Rhomberg et al (2016) noted that the median duration of employment in the cohort was 0.7 years with a median cumulative exposure of 10.7 ppm-years. In addition, analyses by Silver et al. (2002) and Richardson (2008) suggest that the relevant window of exposure for benzene is likely to be no more than 15 -20 years. Richardson et al. (2008) showed that the association between leukemia mortality and benzene exposure was highest in the 10 years immediately after exposure (relative rate per 10 ppm-years = 1.19 [95% CI, 1.101.29]} and reduced to a relative rate of 1.05 per 10 ppm-years (95% CI, 0.971.13) in the period 10 to < 20 years after exposure, with no evidence of association20 years after exposure. Silver et al (2002) also demonstrated evidence of temporal effects showing that the summary relative risk estimate of 13.55 for leukemia in 1961 fell to 2.47 in 1996 (production ended in 1976). Rhomberg et al. (2016) found that when a lag of 15 or 20 years was assumed, many ANLL/AML deaths were observed in the lowest exposure category because their exposures to benzene occurred entirely within 10 years before cancer onset. This observation is consistent with previous findings that leukemia is associated only with benzene exposure within the past 10 to 20 years. These findings taken with those from the study by Rhomberg et al. (2016) suggest that benzene exposure below 4 ppm is not associated with an increased risk of ANLL as exposure that occurred more than 20 years ago does not increase risk, and hence cumulative exposure during the relevant time window will not exceed 80 ppm-years. However, the level may be much closer to 8 ppm as the study by Richardson (2008) suggests that exposures 1020 years previously have much less effect than exposures in previous 10 years.

 

 

Vlaanderen et al. (2010) have attempted to integrate evidence from all available studies that were considered to be of sufficient quality, including the Pliofilm cohort. The shape of the benzeneleukemia exposureresponse curve was assessed using a meta-regression approach. The authors concluded that the natural spline based on all data indicates a significantly increased risk of leukemia [relative risk (RR) = 1.14; 95% confidence interval (CI), 1.041.26] at an exposure level as low as 10 ppm-years. However, several of the studies included in the meta-regression have since been superceded. For instance, the analysis includes a case-control study nested in the Australian Health Watch study (Glass et al. 2003, 2005, 2006) which reported some of the highest risk estimates for exposures below 30 ppm-years. However, Schnatter et al (2012) updated the study as part of a pooled investigation and reported that this association was not replicated in the pooled data, partly because of the absence of a relationship for updated case subjects and partly because of reclassification of some AML case subjects to MDS. The meta-regression also includes estimates from studies by Rushton and Romaniuk (1997) and Schnatter et al. (1996) which form part of the pooled investigation by Schnatter et al (2012). The meta-regression results do appear to be strongly influence by effects at high cumulative exposure and Figure 1B of Vlaanderen et al (2010) suggest little evidence of effect below 50 ppm-years, especially if the findings of the Australian health Watch study, Schnatter et al. (1996) and Rushton and Romaniuk (1997) are replaced by the findings for AML reported by Schnatter et al. (2012). The recent study by Schnatter et al (2012) concluded that MDS may be the more relevant health risk for lower exposures than AML, but also that it is difficult to ascribe precise concentrations of benzene to MDS. Schnatter et al (2012) also reported that peak exposure (at least 1-year employment in jobs likely experiencing >3 ppm exposure for 1560 minutes at least weekly) predicted risk independently of cumulative exposure, but not the reverse. However, a cohort study by Sorahan and Mohammed (2016) has shown no evidence of increased risk in two large cohorts of benzene-exposed workers: 16,467 petroleum distribution workers, and 28,554 oil refinery workers. The UK component of the pooled case-control study conducted by Schnatter et al. (2012) was nested in the cohort of petroleum distribution workers studied by Sorahan and Mohammed (2016). Copley et al. (2017) suggests that exposure to benzene concentrations > 3 ppm may be required to increase the risk of MDS. Nevertheless, the findings from Schnatter et al (2012) and the study by Copley et al. (2017) 

References

Baan R, Grosse Y, Straif K, Secretan B, El Ghissassi F, Bouvard V, Benbrahim-Tallaa L, Guha N, Freeman C, Galichet L and Cogliano V (2009). A review of human carcinogens—Part F: Chemical agents and related occupations. The Lancet Oncology 10, 1143-1144.

Bloeman LJ, Youk A, Bradley TD, Bodner KM & Marsh G (2004). Lymphohaematopoietic cancer risk among chemical workers exposed to benzene. Occup. Environ. Med. 61, 270-274.

Copley et al (2016). Hospital-based case-control study of MDS subtypes and benzene exposure in Shanghai, J Occup Environ Med. 59:349-55.

Cronkite EP, Drew RT, Inoue T and Bullis JE (1985). Benzene hematotoxicity and leukemogenesis. Am J Ind Med 7, 447-456.

Crump KS (1994). Risk of benzene-induced leukemia: a sensitivity analysis of the Pliofilm cohort with additional follow-up and new exposure estimates. J Toxicol Environ Health 42, 219-242.

Crump KS and Allen BC (1984). Quantitative estimates of risk of leukemia from occupational exposure to benzene. Prepared for the Occupational Safety and Health Administration by Science Research Systems, Inc., Ruston, LA. Unpublished.

EC (1993). Occupational exposure limits: Criteria document for benzene. Report EUR 14491 en, ISSN 1018-5593, Commission of the European Communities, pp126.

EC (1999). Council Directive 1999/38/EC of 29 April 1999 amending for the second time Directive 90/394/EEC on the protection of workers from the risks related to exposure to carcinogens at work and extending it to mutagens. Official Journal of the European Communities, L138, 66-69, 1 June 1999.

ECHA (2008). Guidance on information requirements and chemical safety assessment. Chapter R.8: Characterisation of dose [concentration]-response for human health, pp150.

Farris GM, Everitt JI, Irons RD, Popp JA (1993). Carcinogenicity of inhaled benzene in CBA mice. Fund Appl Toxicol 20, 503-507.

Glass DC, Gray CN, Jolley DJ, Gibbons C, Sim MR (2005). Health Watch exposure estimates: do they underestimate benzene exposure? Chem Biol Interact 153-154, 23-32.

Glass DC, Gray CN, Jolley DJ, Gibbons C, Sim MR (2006). The health watch case-control study of leukemia and benzene: the story so far. Ann N Y Acad Sci 1076, 80-9.

Glass DC, Gray CN, Jolley DJ, Gibbons C, Sim MR, Fritschi L, Adams GG, Bisby JA, and Manuell R (2003). Leukemia risk associated with low-level benzene exposure. Epidemiology 14, 569–577. Epidemiology. 2004;15, 509; author reply 510-1.

Goldstein BD, Snyder CA, Laskin S, Bromberg I, Albert RE and Nelson N. (1982). Myelogenous leukaemia in rodents inhaling benzene. Toxicology Letters 13, 169-173.

Gun RT, Pratt N, Ryan P, Roder D (2006). Update of mortality and cancer incidence in the Australian petroleum industry cohort. Occup Environ Med 63, 476-81.

Hayes RB, Yin SN, Dosemeci M, Li GL, Wacholder S, Travis LB, Li CY, Rothman N, Hoover RN and Linet MS (1997). Benzene and the dose-related incidence of haematologic neoplasms in China. Natl Cancer Inst 89, 1065-1071.

IARC (International Agency for Research on Cancer) (1982). IARC monographs on the evaluation of carcinogenic risks of chemicals to humans: some industrial chemicals and dyestuffs. Vol. 29, IARC, pp. 93-148.

Irons RD, Gross SA, Le A, Wang XQ, Chen Y, Ryder J, Schnatter AR (2010). Integrating WHO 2001-2008 criteria for the diagnosis of Myelodysplastic Syndrome (MDS): a case-case analysis of benzene exposure. Chem Biol Interact, 184:30-8.

Koh DH, Jeon HK, Lee SG, Ryu HW (2015). Occup Environ Med. 72:421-7.

Maltoni C, Ciliberti A, Cotti G, Conti B and Belpoggi F (1989). Benzene, an experimental multipotential carcinogen: results of the long-term bioassays performed at the Bologna Institute of Oncology. Environmental Health Perspective 82, 109-124.

NTP (1986). Toxicology and carcinogenesis studies of benzene (CAS No. 71-43-2) in F344/N rats and B6C3F1 mice (gavage studies). NIH publication number 86-2545. Testing laboratory: Battelle Columbus Laboratories. Report no.: TR 289. Study number: NTP TR 289.

Paustenbach DJ, Price PS, Ollison W, Blank C, Jernigan JD, Bass RD, Peterson HD (1992). Reevalution of benzene exposure for the Pliofilm (rubberworker) cohort (1936-1976). J Toxicol Environ Health 36, 177-231.

Paxton MB, Chinchilli VM, Brett SM and Rodricks JV (1994a): Leukaemia risk associated with benzene exposure in the Pliofilm cohort: I. Mortality update and exposure distribution. Risk analysis 14: 147-54.

Paxton MB, Chinchilli VM, Brett SM and Rodricks JV (1994b): Leukaemia risk associated with benzene exposure in the Pliofilm cohort: II. Risk estimates. Risk analysis 14: 155-61.

Richardson DB (2008). Temporal variation in the association between benzene and leukemia mortality. Environ Health Perspect. 116(3):370-4.

Rinsky RA, Young RJ, and Smith AB (1981). Leukaemia in benzene workers. American Journal of Industrial Medicine 2, 217-45.

Rinsky RA, Smith AB, Hornung RW, Filloon TG, Young RJ, Okun AH, and Landrigan PJ (1987). Benzene and leukemia. An epidemiological risk assessment. New England Journal of Medicine 316: 1044-50.

Rinsky RA, Hornung RW, Silver SR and Tseng CY (2002). Benzene exposure and haematopoietic mortality: a long-term epidemiologic risk assessment. Am J Ind Med, Vol 42, pp 474-480.

Rhomberg et al (2016). Evaluation of Acute Nonlymphocytic Leukemia and Its Subtypes With Updated Benzene Exposure and Mortality Estimates: A Lifetable Analysis of the Pliofilm Cohort, J Occup Environ Med. 58(4):414-20. doi: 10.1097/JOM.0000000000000689.

Rushton L and Romaniuk H (1997). A case-control study to investigate the risk of leukaemia associated with exposure to benzene in petroleum marketing and distribution workers in the United Kingdom. Occup Environ Med 54, 152-166.

Schnatter AR, Armstrong TW, Nicolich MJ, et al (1996). Lymphohaematopoietic malignancies and quantitative estimates of exposure to benzene in Canadian petroleum distribution workers. Occup Environ Med, 53:773–781.

Schnatter, AR, Glass DC, Tang G, Irons RD and Rushton L (2012).Myelodysplastic syndrome and benzene exposure among petroleum workers: an international pooled analysis. J Natl Cancer Inst, 104, 1724-37.

Schnatter AR, Rosamilia K, Wojcik NC (2005). Review of the literature on benzene exposure and leukemia subtypes. Chem Biol Interact.153-154, 9-21.

Sorahan and Mohammed (2016). Incidence of Myelodysplastic Syndrome in UK Petroleum Distribution and Oil Refinery Workers, 1995-2011, Int J Environ Res Public Health 13, 474; doi:10.3390/ijerph13050474.

Swaen GM, Scheffers T, de Cock J, Slangen J and Drooge H (2005). Leukemia risk in caprolactam workers exposed to benzene. Ann Epidemiol 15, 21-8.

Swaen GM, van Amelsvoort L, Twisk JJ, Verstraeten E, Slootweg R, Collins JJ, Burns CJ (2010). Low level occupational benzene exposure and hematological parameters. Chem Biol Interact 2010 Jan 14. [Epub ahead of print].

Texas Commission on Environmental Quality (TCEQ) (2007). Development Support Document. Benzene. Chief Engineer’s Office. Available: http://tceq.com/assets/public/implementation/tox/dsd/final/benzene_71-43-2_final_10-15-07.pdf.

Utterback DF and Rinsky RA (1995). Benzene exposure assessment in rubber hydrochloride workers: a critical evaluation of previous estimates. Am J Indust Med 27, 661-76.

Vlaaanderen J, Vermeulen R, Heederik D and Kromhout H (2008). Guidelines to evaluate human observational studies for quantitative risk assessment. Environ Hlth Perspect 116, 1700-1705.

Vlaanderen J, Portengen L, Rothman N, Lan Q, Kromhout H, Vermeulen R. (2010) Flexible meta-regression to assess the shape of the benzene-leukemia exposure-response curve.

Environ Health Perspect, 118:526-32.

WHO (2000). Air quality guidelines for Europe, 2nd ed., World Health Organization Regional Office for 2000 (WHO Regional Publications, European Series No. 91).

Wong O, Harris F, Yiying W, Hua F (2009). A hospital-based case-control study of acute myeloid leukemia in Shanghai: analysis of personal characteristics, lifestyle and environmental risk factors by subtypes of the WHO classification. Regul Toxicol Pharmacol, 55:340-52.