Skip to main content
. 2023 Nov 16;39(5):838–847. doi: 10.1093/ndt/gfad239

Table 4:

Longitudinal associations of BHB between eGFR/year and htTKV/year using linear mixed model analyses.

Crude Model 1 Model 2 Model 3
Variables Est. (95% CI) P-value Est. (95% CI) P-value Est. (95% CI) P-value Est. (95% CI) P-value
eGFR slope R= 0.04, N = 489 R= 0.43, N = 489 R= 0.63, N = 482 R= 0.67, N = 482
 Log2(BHB) 0.35 (0.09 to 0.61) .007 0.30 (0.05 to 0.56) .02 0.40 (0.15 to 0.65) .002 0.33 (0.09 to 0.57) .008
 Sex (female) 0.59 (0.14 to 1.04) .005 –0.01 (–0.48 to 0.45) .9 –0.04 (–0.49 to 0.41) .85
 Age (years) 0.00 (–0.02 to 0.02) .9 0.02 (0.00 to 0.04) .048 0.01 (–0.01 to 0.03) .45
 Log2(copeptin) –0.65 (–0.85 to –0.45) <.001 –0.57 (–0.77 to –0.38) <.001
 SBP –0.03 (–0.05 to –0.02) <.001
PKD2 (ref)*
  PKD1 NT –0.89 (–1.45 to –0.32) 0.002
  PKD1 T –1.19 (–1.75 to –0.63) <.001
  Other/missing 0.06 (–0.74 to 0.85) .9
htTKV growth/year R= 0.03, N = 348 R= 0.16, N = 348 R= 0.22, N = 344 R= 0.26, N = 344
 Log2(BHB) 1.000 (0.995 to 1.005) .9 1.001 (0.997 to 1.006) .55 1.001 (0.996 to 1.006) .74 1.001 (0.996 to 1.006) .67
 Sex (female) 0.983 (0.974 to 0.991) <.001 0.987 (0.978 to 0.996) .004 0.988 (0.979 to 0.996) .006
 Age (years) 1.000 (0.9995 to 1.00) .48 1.000 (0.999 to 1.000) .12 1.000 (0.999 to 1.000) .20
 Log2(copeptin) 1.006 (1.002 to 1.010) .002 1.006 (1.002 to 1.01) .003
 SBP 1.000 (1.000 to 1.001) .13
PKD2 (ref)*
  PKD1 NT 0.997 (0.986 to 1.008) .63
  PKD1 T 1.001 (0.990 to 1.012) .89
  Other/missing 0.975 (0.960 to 0.989) .001

The estimates and P-values were calculated using linear mixed model analyses. The estimates are the variables and their interaction with time which is the effect of the variables on eGFR slope (mL/min/1.73 m2 per year) (top part of the table) or htTKV growth per year (bottom part of the table).

The independent variables are baseline log2(BHB) (crude), adjusted for sex, age (Model 1), additionally adjusted for log2(copeptin) (Model 2) and additionally adjusted for SBP and PKD mutations (Model 3). The estimates for dependent variable htTKV growth are back-log transformed. For categorical variables, estimates show fold changes compared with the reference category. For continuous variables, the estimate indicates the fold change per one-unit increase. When the independent variable is log2-transformed, a doubling (one unit increase) of the independent variable corresponds to the fold change according to the estimate. The fold change can be calculated to percentage difference = (fold change – 1)*100%.

*PKD mutation was used as dummy variable with PKD2 as reference group.

Est., estimate; PKD, polycystic kidney disease; NT, non-truncating; T, truncating.