Fig 1.
Baseline phosphorus and time-averaged phosphorus in the follow-up.
a) The distribution of the baseline phosphorus and TA-P. The number of patients decreased below 3.5 mg/dL of serum phosphorus while the percentage dramatically increased in the range of 3.50–3.59 mg/dL of time-average phosphorus. Hyperphosphatemia accounted for 2.6% and 3.7% at baseline and in the follow-up, respectively. b) The relationship between TA-P and baseline phosphorus. A regression line is depicted in black line and a unity of the slope is overlapped in a dotted line. The correlation coefficient was 0.69 (p < 0.001). Two lines converge at 3.9 mg/dL of phosphorus. TA-P increased below baseline phosphorus 3.5 mg/dL and decreased above baseline phosphorus 3.5 mg/dL due to the remarked increase in distribution as shown in a). Abbreviation: TA-P, time-averaged phosphorus.
Table 1.
Clinical characteristics of the cohort (n = 803).
Fig 2.
Hazard ratios at the different thresholds of serum phosphorus in the follow-up.
Hazard ratios above versus below the indicated threshold of TA-P was obtained by applying standard Cox proportional hazards models. The total patients (n = 803) were separated at each threshold followed by survival analysis for ESRD as outcome (n = 110). Univariate analysis was performed for a threshold time-average phosphorus as a sole independent covariate. Multivariate analysis was adjusted for the baseline covariates including sex, age, diabetic nephropathy, estimated GFR, albumin, Na-Cl, phosphorus, LDL-C and proteinuria. The multivariate regression analysis showed the significant hazard ratio of higher TA-P up to 4.4 mg/dL. Abbreviation: TA-P, time-averaged phosphorus.
Fig 3.
Hazard ratios at the different thresholds of serum phosphorus in the follow-up by propensity score analysis.
Hazard ratios (HRs) below and above the indicated threshold of TA-P by applying stratified Cox proportional hazards models. The total patients (n = 803) were trimmed by removing the participants with non-overlapped propensity scores at each threshold. The numbers of test participants and ESRD are shown below the graph. Univariate analysis was performed for TA-P as a sole independent covariate. Multivariate analysis was adjusted for the baseline covariates including sex, age, diabetic nephropathy, estimated GFR, albumin, Na-Cl, phosphorus, LDL-C and proteinuria. The multivariate regression analysis showed the significant HR of higher TA-P up to 4.2 mg/dL. Abbreviation: TA-P, time-averaged phosphorus.
Fig 4.
Standardized differences of baseline covariates before and after propensity score matching with different thresholds of serum phosphorus in the follow-up.
a) 3.4 mg/dL, b) 3.6 mg/dL, c) 3.8 mg/dL, d) 4.0 mg/dL of serum phosphorus in the follow-up. Standardized differences below 0.1 are recognized as good balance between two groups. Covariates which did not reach below 0.1 is indicated by asterisk (*). Abbreviations: TA-P, time-averaged phosphorus; eGFR, estimated glomerular filtration rate; DMN, diabetic nephropathy; BMI, body mass index; SBP, systolic blood pressure; Hb, hemoglobin; WBC, white blood cell; Plt, platelet; Alb, albumin; UA, uric acid; Na, sodium; K, potassium; Cl, chloride; cCa, albumin-corrected calcium; P, phosphorus; CRP, C reactive protein; LDL-C, low-density lipoprotein cholesterol; TPU/CrU, urine total protein divided by urine creatinine; UB_score, urine blood score; RASi, RAS inhibitor.
Fig 5.
Kaplan-Meier plots separated by different thresholds of serum phosphorus in the follow-up after the propensity score matching.
a) 3.4 mg/dL, b) 3.6 mg/dL, c) 3.8 mg/dL, d) 4.0 mg/dL of serum phosphorus in the follow-up. The risk tables are shown below the graph and p values are computed by stratified log-rank test. Hazard ratios of the incidence of ESRD regarding the thresholds of serum phosphorus in the follow-up were 12.83 (95% CI 6.12–21.92), 3.43 (95% CI 1.75–6.71), 2.56 (95% CI 1.54–4.26) and 2.29 (95% CI 1.39–3.79), respectively (See Table 2).
Table 2.
Kaplan-Meier analysis before and after propensity score matching, and hazard ratio, absolute risk reduction and number needed to treat over 5 years of follow-up.