Table 2.

General linear mixed effects regression model of the association between the individual components of BP (systolic BP, diastolic BP, and pulse pressure) and eGFR for splines of time

VariableEstimate (Change in eGFR)P ValueVariableEstimate (Change in eGFR)P ValueVariableEstimate (Change in eGFR)P Value
Intercept15.54<0.001Intercept15.34<0.001Intercept15.69<0.001
SBP−0.02<0.001DBP−0.06<0.001PP−0.010.24
SBP×1–48 d0.140.004DBP×1–48 d0.32<0.001PP×1–48 d0.050.34
SBP×49–168 d−6.970.02DBP×49–168 d13.080.01PP×49–168 d−3.560.26
SBP×169–411 d10.220.02DBP×169–411 d18.490.02PP×169–411 d5.510.24
SBP×411–952 d−3.500.04DBP×411–952 d−5.430.06PP×411–952 d−2.270.20
P value for nonlinearitya<0.001<0.001<0.001
P value for interactionb<0.001<0.0010.49
  • Models are adjusted for age at cohort entry, sex, and baseline comorbidities (coronary artery disease, congestive heart failure, malignancy, hypertension, peripheral vascular disease, diabetes, and angiotensin-converting enzyme inhibitor/angiotensin receptor blocker use) as well as repeat measures of hemoglobin, albumin, phosphate, potassium, bicarbonate, and proteinuria. eGFR is in milliliters per minute per 1.73 m2. SBP, systolic BP; DBP, diastolic BP; PP, pulse pressure.

  • a P values represent statistical significance testing of likelihood ratios for models containing nonlinearity terms for time compared with linearity terms for time.

  • b P values represent statistical significance testing of likelihood ratios for models containing terms for interactions with BP and time compared with no interactions.