ppm) with those of 268 nonexposed workers from the same factory. The study found no abnormalities in white-cell count, red-cell count, hemoglobin concentration, platelet count, or mean red-cell volume. The study did find that smoking affected many hematologic parameters, underlining the importance of controlling for confounding. Many other studies of the effects of benzene exposure on hematologic parameters that were reviewed provided no information on persistent long-term health effects (e.g., Aksoy et al., 1971; Bogadi-Sare et al., 1995, 1997; Collins et al., 1997; Khuder et al., 1999; Rothman et al., 1996; Tsai et al., 1983; Ward et al., 1996).
Only a few studies have examined aplastic anemia in relation to exposure to other specific solvents or to solvents in general. Several of the insecticide-exposure studies described above also examined exposure to solvents. The population-based, case-control study in Thailand (Issaragrisil et al., 1996) compared 284 cases of aplastic anemia identified in 40 hospitals in Bangkok and 15 hospitals in rural areas. The study enrolled four hospital controls of similar age and sex for each case. Two hematologists confirmed the diagnoses of aplastic anemia. Using interviews with the case and control subjects, the researchers examined several risk factors for aplastic anemia. For the cases and controls drawn from Bangkok hospitals, there was a strong association with a history of solvent exposure (RR=4.6, 95% CI=2.5–8.7). About 40% of the total cases came from rural hospitals, and no association was noted when those cases were compared with their controls. The study reported positive associations for other risk factors (such as grain farming, hepatitis A, and low socioeconomic status) and multivariate analyses adjusted for the many possible confounders. However, the study presents little information on exposure-assessment methods, participation rates, or the conduct of the interviews. Also, such hospital-based case-control studies are vulnerable to selection and recall bias.
Using the French national register of aplastic anemia, Guiguet and colleagues (1995) studied 98 patients with aplastic anemia (recorded in the register in 1985–1988) and two groups of controls: 181 selected from the same hospital as the cases and 72 referred by case patients from among neighbors. Interviews were conducted to determine occupational and medical histories, and a toxicologist coded the occupational exposures (any exposure or a “large level of exposure”). The study reported no association between aplastic anemia and exposure to all types of solvents compared with hospital controls (OR=0.9, 95% CI=0.5–1.7) or neighbor controls (OR=0.6, 95% CI=0.3–1.4). Analysis of exposure to various classes of solvents revealed no consistently increased risk; for example, for higher exposure to halogenated solvents, the OR was 1.3 (95% CI=0.6–2.7) compared with hospital controls. Although the study was limited by a 50% participation rate, interviews were conducted at diagnosis, and the investigators found no evidence of participation bias in the case group.
The case-control study in Baltimore, Maryland, described above (Linet et al., 1989) found an association of aplastic anemia with self-reported exposure to paint (OR=6.1, 95% CI=1.2–29.7), but there was no association with the occupation of painter. The study reported a slightly elevated risk for aplastic anemia and exposure to any solvents (OR=1.1, 95% CI=0.5–2.7). However, as noted above, the inferences from this study are limited by the fact that 41% of the patients with aplastic anemia were under 20 years old at diagnosis and would not have experienced substantial occupational exposure.