Epidemiological Issues in Determining Whether Benzene Causes Lymphatic

Epidemiological Issues in Determining Whether Benzene Causes Lymphatic Cancer or A Toxicologist’s Defense Against the Pump Handle Bernard D Goldstein University of Pittsburgh Graduate School of Public Health

The Three Laws of Toxicology • The dose makes the poison • Chemicals have specific effects • Humans are animals

Questions for Discussion • What does the epidemiology literature tell us about whether benzene causes non-Hodgkin’s lymphoma (NHL) or multiple myeloma (MM)? • Can epidemiology tell us whether benzene exposure doubles the risk of NHL or MM - and why should anyone care about the doubling of risk?

The National Safety Council Congress…. . In 1921 held a session on benzene poisoning. Dr. Lothar E. Weber of the Boston India Rubber Laboratory stated that benzene “has been criticized as very dangerous (and) very injurious… and, personally, I feel an injustice has been done to this particular substance. ”… In response, C. F. Horan of the Hood Rubber Company replied that inhalation experiments with benzene, toluene, and xylene on guinea pigs and rabbits, showed acute toxicity of benzene compared to the other two compounds. Not persuaded, Weber rejoined that he was not going to change his opinion altogether on the basis of a few guinea pig experiments. Hounshell & Smith, 1988

Benzene and Alkylbenzenes CH 3 Benzene CH 2 Toluene CH 3 Ethylbenzene H 3 C CH 3 C Xylene H Cumene

Occam’s Razor is Dull Simplest Proposition: One metabolite acting through one mechanism attacking one target Likely Truth: Multiple metabolites acting through multiple mechanisms attacking multiple targets

Hematologic Effects of Benzene Causality Proven – Aplastic Anemia – Myelodysplasia – Acute Myelogenous leukemia (Including Acute Myelomonocytic Leukemia, Acute Promyelocytic Leukemia, Erythroleukemia)

Evidence Supporting Benzene Leukemogenesis 1. Biomedical Plausibility 2. Case Studies 3. Epidemiology A. Numerator Specific B. Denominator Specific

Carcinogenic potency of benzene calculated on the basis of animal data Data Base Lifetime risk per ppm per ug/m 3 Female rats (Maltoni, et. al, 1982) a 3. 4 x 10 – 2 1. 1 x 10 – 5 Male rats (NTP, 1984) a 2. 0 x 10 – 2 6. 0 x 10 – 6 Female rats (NTP, 1982) a 3. 3 x 10 – 2 1. 0 x 10 – 5 Male rats (Snyder, et. al, 1980) b 1. 4 x 10 – 2 4. 3 x 10 – 6 2. 4 x 10 – 2 7. 3 x 10 – 6 GEOMETRIC MEAN

Observed and Expected Deaths due to Leukemia in British Male Oil Refinery Workers Observed Expected O/E All deaths 4, 406 5, 260 0. 84 All neoplasms 1, 147 1, 286 0. 89 30 32 0. 94 Leukemia Alderson & Rushton, 1982

Case Control Study of Benzene Exposure and Leukemia in 36 British Male Oil Refinery Workers Benzene Controls Cases Exposure Low 18 72 Medium or High 18 36 RR (95% CI) 2. 0 (0. 93 – 4. 30) Rushton & Alderson, 1981

Case Control Study of Benzene Exposure and Leukemia in 36 British Male Oil Refinery Workers Logistic Models Matched on Year of Birth Benzene Exposure Relative Risk Confidence Interval Benzene exposure 2. 01 0. 94 – 4. 28 Benzene exposure plus year of entry 2. 26 1. 01 – 1. 43 Benzene exposure plus length of service 2. 99 1. 24 – 7. 20 Rushton & Alderson, 1981

Hematologic Effects of Benzene Causality Probable but Unproven – – – Acute Lymphatic Leukemia Non-Hodgkin’s Lymphoma Multiple Myeloma Paroxysmal Nocturnal Hemoglobinuria Chronic Myelogenous Leukemia

Hematologic Effects of Benzene Causality Possible – Hodgkin’s Disease – Chronic Lymphocytic Leukemia (and other Myeloproliferative Disorders)

Multiple Myeloma • Plasma cell tumor, usually of bone marrow • Plasma cells are related to B Lymphocytes and have the function of producing antibody • Diagnosis usually made based upon the presence of a monoclonal protein spike on serum protein electrophoresis, and on the presence a large numbers of plasma cells in the bone marrow.

Normal Monoclonal Gammopathy

Non-Hodgkin’s Lymphoma • Lymphocytic tumors diagnosed by exclusion - not Hodgkin’s disease nor lymphocytic leukemias • Broad and overlapping range of disease entities and etiologies. • Immune suppression common to a number of causative factors, including HIV infection.

Biological Plausibility of Causal Relationship of Benzene to Multiple Myeloma • Multiple myeloma is a tumor of plasma cells which are a form of B lymphocytes • Exposure to benzene destroys B lymphocytes and causes chromosomal abnormalities in B lymphocytes • Benzene is a known cause of leukemia, a bone marrow cancer, through a mechanism that leads to the presence of a carcinogenic metabolite within the bone marrow. • Multiple myeloma is a bone marrow tumor.

Role of Biological Plausibility in Determining Causal Relations of Benzene to Multiple Myeloma • Benzene causes the formation of a carcinogen that is specific to the organ at risk and that affects the basic cell type, including producing cytogenetic abnormalities.

Biological Plausibility of Causal Relationship of Benzene to Non-Hodgkin’s Lymphoma • Non-Hodgkin’s Lymphoma is a lymphocytic tumor • Exposure to benzene destroys lymphocytes and causes chromosomal abnormalities in lymphocytes • Benzene is a known cause of leukemia, a bone marrow cancer, through a mechanism that leads to the presence of a carcinogenic metabolite within the bone marrow. • The bone marrow is a lymphoid organ. • Rats exposed to benzene develop lymphomas

Pluripotential Bone Marrow Stem Cell(s) Matures to precursors of: • • Red blood cells Platelets Granulocytic white blood cells Lymphocytic white blood cells

Case control study • 309 matched pairs of hematopoietic and lymphoid neoplasms in Kanawha County, WV. • “association between chemical industry work and death due to non-Hodgkin’s lymphoma, multiple myeloma, and lymphoid leukemia…” • For NHL, OR 3. 11, p =. 003 for those who died at age <65 • For MM, OR 2. 39, p =. 039 for all age groups • For all hematopoietic and lymphoid neoplasms; OR 3. 31, p =. 001

Case control study • 309 matched pairs of hematopoietic and lymphoid neoplasms in Kanawha County, WV. • “association between chemical industry work and death due to non-Hodgkin’s lymphoma, multiple myeloma, and lymphoid leukemia…” • For NHL, OR 3. 11, p =. 003 for those who died at age <65 • For MM, OR 2. 39, p =. 039 for all age groups • For all hematopoietic and lymphoid neoplasms; OR 3. 31, p =. 001 • Massoudi, Talbott, Day, Swerdlow, Marsh and Kuller. Amer J Indust Med, 1997