Table 3

Multivariate analysis of subtype of myeloid and lymphoid leukaemia cases by job exposure matrix

ExposureAcute myeloid leukaemia n=113Chronic myeloid leukaemia n=46Chronic lymphoid leukaemia n=216
Multivariate HR (95% CI)p Value*Multivariate HR (95% CI)p Value*Multivariate HR (95% CI)p Value*
Pesticides
 No1.000.791.000.341.000.86
 Low0.85 (0.41 to 1.79)2.01 (0.81 to 4.99)1.05 (0.65 to 1.69)
 High1.19 (0.60 to 2.37)1.33 (0.50 to 3.56)1.03 (0.63 to 1.69)
Herbicides
 No1.000.911.000.351.000.89
 Low
 High0.97 (0.54 to 1.73)1.45 (0.66 to 3.20)1.02 (0.69 to 1.53)
Insecticides
 No1.000.791.000.341.000.86
 Low0.85 (0.40 to 1.79)2.01 (0.81 to 4.99)1.05 (0.65 to 1.69)
 High1.19 (0.60 to 2.37)1.33 (0.50 to 3.56)1.03 (0.63 to 1.69)
Aromatic solvents
 No1.000.171.000.411.000.52
 Low1.15 (0.73 to 1.80)0.67 (0.33 to 1.35)0.97 (0.72 to 1.32)
 High1.83 (0.86 to 3.88)0.97 (0.23 to 4.13)0.71 (0.33 to 1.53)
Benzene
 No1.000.281.000.301.000.37
 Low1.06 (0.63 to 1.81)1.00 (0.45 to 2.22)1.11 (0.78 to 1.58)
 High1.52 (0.78 to 2.98)1.97 (0.75 to 5.19)0.56 (0.27 to 1.14)
Chlorinated solvents
 No1.000.581.000.831.000.77
 Low1.20 (0.79 to 1.80)0.73 (0.38 to 1.41)0.97 (0.73 to 1.28)
 High0.79 (0.19 to 3.30)1.74 (0.50 to 6.10)0.91 (0.37 to 2.26)
Trichloroethylene
 No1.000.771.000.801.000.33
 Low1.00 (0.66 to 1.51)0.98 (0.51 to 1.87)0.93 (0.69 to 1.24)
 High0.73 (0.22 to 2.38)0.77 (0.18 to 3.42)0.64 (0.28 to 1.47)
Formaldehyde
 No1.000.741.000.701.000.36
 Low1.01 (0.65 to 1.57)0.92 (0.46 to 1.84)1.12 (0.83 to 1.52)
 High1.45 (0.46 to 4.56)
Metals
 No1.000.981.000.971.000.34
 Low0.97 (0.59 to 1.60)0.82 (0.36 to 1.84)0.89 (0.63 to 1.26)
 High1.06 (0.45 to 2.54)1.23 (0.41 to 3.72)0.75 (0.38 to 1.50)
Contact with animals or animal products
 No1.000.451.000.281.000.94
 Low0.46 (0.11 to 1.86)2.43 (0.85 to 7.01)1.05 (0.54 to 2.07)
 High0.85 (0.48 to 1.51)1.37 (0.62 to 3.02)1.01 (0.69 to 1.49)
Ionising radiation
 No1.000.511.000.541.000.14
 Low0.83 (0.40 to 1.72)0.73 (0.22 to 2.39)1.45 (0.96 to 2.17)
 High
  • Multivariate analyses were adjusted for sex, smoking status, alcohol intake, age at recruitment and country.

  • *Tests of linear trend across categories were conducted by fitting a model treating the different tertile categories as a single ordinal variable.