PSTR | Coefficient (95% Confidence Intervals) | P Value |
---|---|---|

Difference between biocompatible and standard solutions at start of PD | −0.049 (−0.09 to −0.01) | 0.02 |

Effect of time with standard solutions, yr | Linear coefficient −0.028 (−0.04 to −0.01); quadratic coefficient 0.0009 (>−0.01 to <0.01) | 0.001^{a} |

Effect of time with biocompatible solutions, yr | Linear coefficient 0.061 (0.03 to 0.10); quadratic coefficient −0.010 (−0.02 to >−0.01) | 0.003^{a} |

Center effect | −0.13 to <0.01 | 0.02 |

Icodextrin use | 0.045 (0.03 to 0.06) | <0.001 |

Average dextrose concentration in 24 h, % | −0.022 (−0.05 to <0.01) | 0.06 |

Average dextrose concentration with time interaction | 0.015 (<0.01 to 0.02) | <0.001 |

Baseline dialysate IL-6 (log transformed), pg/ml | 0.048 (0.03 to 0.07) | <0.001 |

Urine volume, L | 0.011 (0.00 to 0.02) | 0.04 |

Peritonitis with biocompatible solutions | −0.014 (−0.03 to 0.01) | 0.11 |

Peritonitis with standard solutions | 0.020 (0.01 to 0.03) | <0.001 |

Multilevel multivariable model showing the effect of covariates on PSTR. Coefficients show the adjusted rise in PSTR for a 1-unit change in covariate. For 1 L of urine volume, there is an increase of 0.011. For an increase in time of 2 yr, the change represents the sum of the two coefficients for the solution type (

*e.g.*, standard solutions have a change of 0.0009×2×2–0.028×2). At the start of PD, biocompatible solution use was associated with slower PSTR, which was increasing significantly over the next 2–3 yr and then flattened off; for standard solutions, a trivial reduction in PSTR and then a slight rise were seen in the model (negative linear and positive quadratic time coefficients). Independent of these trends, use of icodextrin, baseline inflammation, and peritonitis were all associated with faster PSTR, the latter only when exposed to standard solutions. Higher dialysate dextrose concentration was associated with faster PSTR after a prolonged period on PD (positive interaction with time), but there was little association at the start. Covariates not included in the final model (because they were not significant) include APD/CAPD use, plasma IL-6, sex, age, and comorbidity status. Center had a significant effect (range of coefficient, −0.13 to 0.001). PSTR, peritoneal solute transport rate; PD, peritoneal dialysis.↵a

*P*values for overall effect from Wald statistic.