A Critique of the Exposure Assessment in the Epidemiologic Study of Benzene-Exposed Workers in China Conducted by the Chinese Academy of Preventive Medicine and the US National Cancer Institute

Regul Toxicol Pharmacol. 1999 Dec;30(3):259-67. doi: 10.1006/rtph.1999.1354.


As reviewed in some detail in the present paper, workers employed in a wide variety of industries were included in the Chinese benzene study, and were exposed to not only benzene but also a wide range of other industrial chemicals. To attribute any or all health effects observed in the exposed cohort to benzene without examining other concomitant exposures is not appropriate. Although it was stated that one of the major objectives of the expanded study was to examine the effects of other risk factors, no such examination was made in any of the analyses in the expanded CAPM-NCI study. The CAPM-NCI study suffered from a number of limitations. One of the most serious limitations of the study involved the exposure estimates developed by the US NCI team. Comparing the assumptions used in the development of estimates and the exposure estimates themselves to actual data reported previously by the Chinese investigators revealed numerous inconsistencies and, in many cases, large discrepancies. It appeared that the exposure estimates were consistently lower than the actual exposure data. The so-called indirect validation conducted by the NCI team served no useful purpose, since by definition it could not validate the absolute values of the estimates. NCI was fully aware of some of the inadequacies of its exposure estimates. Although in a 1994 paper, the NCI team recognized that little confidence could be attached to the estimated (e.g., only 2% of the estimates for the time interval 1949-1959 and only 6% of the estimates prior to 1975 were rated in the high confidence category), the inadequacy of the estimates was never mentioned or discussed in any subsequent analyses or in the latest report (Hayes et al., 1998). Instead, the exposure of the workers was hailed as "well characterized" (Hayes et al., 1998). In conclusion both CAPM and NCI have made substantial efforts in studying the relationship between benzene exposure and various malignancies. Unfortunately, there were many inherent problems in the data as well as serious limitations in the exposure estimates. Because of these unresolved problems and limitations, many of the results in the CAPM-NCI study are unreliable. Therefore, the conclusions of the study, particularly those involving exposure estimates, are not justified.

Publication types

  • Comparative Study

MeSH terms

  • Benzene / adverse effects*
  • Benzene / poisoning
  • China / epidemiology
  • Dose-Response Relationship, Drug
  • Humans
  • Leukemia / chemically induced*
  • Leukemia / epidemiology
  • Occupational Diseases / chemically induced*
  • Occupational Diseases / epidemiology
  • Occupational Exposure*


  • Benzene