Table 5

Results for log serum glucose (fasting glucose, mmol/l), linear regression

Study Coefficient for TCDD (95% CI)5-150 Coefficient for log TCDD
(95% CI)5-150 		Change in log serum glucose by quartile of TCDD >10 ppt (95% CI), referent <=10 ppt5-151
Ranch Hand 0.00006 (−0.00009 to 0.00021) 0.0016 (−0.0025 to 0.0057)−0.001 (−0.016 to 0.016), −0.001 (−0.017 to 0.014), −0.011 (−0.028 to 0.004), 0.014 (−0.024 to 0.038)
NIOSH 0.00004 (−0.00001 to 0.00009) 0.0033 (−0.0047 to 0.0113)−0.016 (−0.077 to 0.045), −0.017 (−0.065 to 0.031), −0.033 (0.078 to 0.012), 0.014 (−0.017 to 0.045)
Combined 0.00004 (0.00001 to 0.00008) 0.0025 (−0.0010 to 0.0060)−0.004 (−0.019 to 0.013), −0.005 (−0.016 to 0.011), −0.017 (−0.033 to 0.007), 0.018 (0.001 to 0.035)
R 2 combined model 0.09 0.09 0.09
Combined without the three most influential observations−0.00001 (−0.00005 to 0.00003)−0.0004 (−0.0039 to 0.0032)−0.003 (−0.018 to 0.013), −0.005 (−0.020 to 0.011) −0.017 (−0.033 to −0.001), 0.004 (−0.012 to 0.020)

  • ↵5-150 A test for heterogeneity of exposure effect between studies gave p=0.90 for TCDD and p=0.24 for log TCDD, when data included three most influential observations.

  • ↵5-151 Quartiles for each study and for combined data are 10–17, 17–30, 30–78, ⩾78 ppt for those >10 ppt.