Fig 1.
Serum FGF23 was negatively associated with the eGFR, but not the UACR, in patients with T1D.
(A) The distribution of serum full-length FGF23 levels. The median (Q1, Q3) serum FGF23 level was 36.6 (30.7, 47.1) pg/ml. (B and C) Descending quartiles of the eGFR were associated with significantly increased mean (±SE) serum FGF23 levels (P for linear trend = 0.0371) (B). On the other hand, ascending quartiles of UACR were not significantly associated with increased mean (±SE) serum FGF23 level (P for linear trend = 0.209) (C). The Jonckheere-Terpstra trend test was used in these analyses. (D-F) The associations between the serum FGF23 level and the eGFR (r = -0.292, P = 0.0016) (D), UACR (r = 0.116, P = 0.212) (E), serum Pi level (r = 0.273, P = 0.0027) (F). Non-normally distributed variables were subjected to natural logarithmic transformation. Pearson’s correlation coefficients were used in these analyses.
Table 1.
Clinical characteristics of the subjects.
Table 2.
Associations between FGF23 and each parameter.
Table 3.
Association between FGF23 and eGFR in patients with type 1 diabetes.