Application of a latent variable model for a multicenter study on early effects due to mercury exposure

Roberto Lucchini, Stefano Calza, Donatella Camerino, Plinio Carta, Adriano Decarli, Giovanni Parrinello, Leonardo Soleo, Roberto Zefferino, Lorenzo Alessio

Research output: Contribution to journalArticlepeer-review


A latent variable model was applied to the results of an Italian multicenter nation-wide cross-sectional study to assess the earliest health effects due to mercury (Hg) exposure caused by occupation, dental amalgams (DENTAM) and fish eating (FISH). The studied population included subjects recruited from four different geographical areas. A total number of 122 workers from chloroalkali plants and production of thermometers and neon lamps formed the occupationally exposed group, whereas 196 subjects, recruited from the same areas and not occupationally exposed to mercury, formed the control group. Neuropsychological functions were assessed with neurobehavioral testing including vigilance, motor and cognitive function, tremor measurements, and with symptoms concerning neuropsychological and mood assessment. Neuroendocrine function was examined with the measurement of prolactin (PRL) level. Parameters of immunological and renal function were also measured. The target population was characterized by the number and surface of dental amalgams and consumption of fish. In the exposed workers the average urinary mercury (U-Hg) was 10.4 ± 6.9 (geometric mean 8.3, range 0.2-35.2) μg/g creatinine, whereas in the control group it was 1.9 ± 2.8 (geometric mean 1.2, range 0.1-33.2) μg/g creatinine. The preliminary results indicated that finger tapping (FT) and the Branches alternate movement task (BAMT) coordination test were associated with the occupational exposure (OCCEXP). PRL was significantly decreased among the exposed workers, and inversely related to U-Hg. Among the immunological and renal parameters, cytokine serum interleuchin-8 (sIL8) and β2 micro globulin (β2MG) were lower in the exposed group and negatively correlated to U-Hg. Small-size fish consumption was associated to a beneficial effect on symptoms reporting. No effects were observed concerning dental amalgams. After first evaluating the relationship between mercury exposure and each indicator of effect, further assessment was performed to identify the earliest effects related to mercury exposure among those who resulted in being associated in the preliminary elaboration. Two latent variables exposure and effect were identified, integrating respectively the different forms of exposure (occupational, due to dental amalgams and fish consumption) and the indicators of effects (FT, BAMT, PRL, sIL8, β 2MG). Confounding factors (age, alcohol, body mass index (BMI)) were considered in the same model. This further analysis showed that an inverse association of occupational exposure to mercury with PRL and BAMT, with Hg-U mediating the effect on PRL, was predominant with respect to the other form of mercury exposure, the other indicators of effect and the confounders. In conclusion, this study supports the finding of alterations of neuroendocrine secretion and motor coordination at very low occupational exposure levels of inorganic mercury, below the current ACGIH Biological Exposure Index. These changes occur at lower levels than other subtle effects on the renal function and the immunitary system. On the contrary, dental amalgams and small-size fish consumption do not seem to be associated to any adverse health effect at these exposure levels.

Original languageEnglish
Pages (from-to)605-616
Number of pages12
Issue number4-5
Publication statusPublished - Aug 2003


  • Dental amalgams
  • Fish eating
  • Inorganic and organic mercury
  • Neurotoxicity
  • Occupational exposure

ASJC Scopus subject areas

  • Cellular and Molecular Neuroscience
  • Neuroscience(all)
  • Toxicology


Dive into the research topics of 'Application of a latent variable model for a multicenter study on early effects due to mercury exposure'. Together they form a unique fingerprint.

Cite this