Summary of multiple multivariate logistic regressions for nephropathy

Covariate effects from multiple modelsdfχ2P valueR2
Unadjusted effect
    DCCT mean A1C110.40.00120.062
    EDIC study mean A1C132.0<0.00010.189
    Log DCCT close-out AER113.60.00020.081
    Furosine, CML, pentosidine, fluorescence, acid/pepsin soluble619.30.00360.114
    Pentosidine, fluorescence, acid/pepsin soluble412.60.01350.075
    Furosine, CML218.20.00010.108
Adjusted effect
    DCCT mean A1C effect adjusted for
Furosine, CML10.00.96430.000
    DCCT mean A1C, log AER effect adjusted for
Furosine, CML210.80.00460.064
    EDIC study mean A1C effect adjusted for
Furosine, CML123.1<0.00010.137
    Furosine, CML effect adjusted for
Pentosidine, fluorescence, acid/pepsin soluble211.00.00410.065
DCCT mean A1C, log AER213.10.00140.078
DCCT mean A1C212.80.00160.076
EDIC study mean A1C214.40.00080.085
    Pentosidine, fluorescence, acid/pepsin soluble effect adjusted for
Furosine, CML44.50.33940.027
    Furosine effect adjusted for
DCCT mean A1C17.10.00790.042
EDIC study mean A1C17.10.00760.042
    CML effect adjusted for
DCCT mean A1C16.30.01230.037
EDIC study mean A1C110.70.00110.064
  • Dependent variable: development of microalbuminuria or worse from DCCT close out through EDIC study year 10. As defined in the statistical methods section, χ2 and P values are from likelihood ratio test, and R2 is entropy R2.