Author's response to reviews Title:Urinary phosphorus excretion per creatinine clearance as a prognostic marker for progression of chronic kidney disease: A retrospective cohort study Authors:

1) Lines 6-8: How does the ratio of U-P/24h-CCr reflect the phosphate loss for each nephron? A proof that 24h-CCr is a marker for nephron number is missing. Why is it important to normalize to creatinine clearance? In animal experience, renal biopsy and magnetic resonance imaging were used to estimate glomerular number. Fulladosa et al reported that a positive correlation between GFR and Glomerular number in stable renal transplants. (Fulladosa X, Moreso F, Naráez JA, Grinyó JM, Serón D. Estimation of total glomerular number in stable renal transplants. J Am Soc Nephrol 2003; 14: 2662-8). Because 24h-CCr can be measured less invasively, we used 24h-CCr as marker for nephron number. We have therefore added the following text as part of introduction (p6-7 line 12-1). In animal experience, renal biopsy and magnetic resonance imaging were used to estimate glomerular number. Previous study reported that a positive correlation between GFR and Glomerular number in stable renal transplants. (11). Therefore, we used 24h-CCr as marker for nephron number. We have also added the following reference. 11. Fulladosa X, Moreso F, Naráez JA, Grinyó JM, Serón D. Estimation of total glomerular number in stable renal transplants. J Am Soc Nephrol 2003; 14: 2662-8 2) Many studies deal with negative outcome of high levels of phosphate in serum. What is the advantage of this established marker than for example phosphate in serum? Can you show a graph that it is an earlier marker than phosphate in serum in these patients by a graph (results section)? We have added new the figure that shows U-P/CCr elevated earlier than serum phosphorus in CKD as reviewer indicated. 3) Usually FGF23 is one marker which increases early in serum. How is the association of U-P/24-CCr with FGF 23 in serum? FGF23 is a hormone that increases phosphate excretion per nephron by increasing the fractional excretion of phosphate and increasing GFR. So we have assumed there is a positive correlation between 24h-CCr and FGF23 in serum 4) What is the key result for the conclusion? Material & Methods: The key result for the conclusion is that higher U-P/24h-CCr showed a higher risk for the composite outcomes. 5) The source of protein is of importance, for example for association of phosphate with protein rich foods has reduced bioavailability from plant origin. Phosphate intake is not mentioned but excretion is highly dependent on phosphate intake. In our study, the fixed diet regimen was served for the patient. The proportion of animal protein in the diet regimen was 50-60 percent of total. Therefore phosphate intake of each patient is fixed about 0.8 to 1.0 g / kg, standard body weight. We have added the following text as reviewer indicated. The proportion of animal protein in the diet regimen was 50-60 percent of total. Therefore, phosphate intake of each patient is also fixed about 0.8 to 1.0 g / kg, standard body weight. 6) Table 1: An average BMI of 59.8 is highly obeseis that the correct figure? Please give a more detailed description of CKD patients in the results section. That is not the correct figure. Table 1 have been changed as reviewer detected. 7) Table 1: 86% of the patients were in CKD stage 5. Were all CKD stages analysed all together? Was there an adjustment for CKD stages? There was not an adjustment for CKD stages but for eGFR, because highly significant statistical correlation between CKD stage and eGFR was observed. 8) How war the tubular reabsorption calculated? Usually it considers GFR and Creatinine. Tubular absorption of phosphate (TRP) was calculated using the formula as follows: TRP = 1(U-phosphate × S-creat) / ( S-phosphate × U-creat) We have added the following text. TRP was calculated using the following formula: TRP = 1(Urinary phosphate × Serum creatinine) / ( Serum phosphate × Urinary creatinine). Results: 9) Figure1 a) and b): The impact of showing the regression of both approaches is not clear. Graphs need more detailed descriptions. Are all CKD stages included? Figure 1 shows assessing the assumption in Cox regression. Hazard ratio in each patient was calculated, and then patients were divided into two groups in the median of hazard ratio. The graph of the log(-log(survival)) versus log of survival time graph should resulted in parallel curves if the Cox regression models was acceptable. We have change the following figure legend (Figure 1.) Figure 1. Proportional hazards with the applied predictors in the Cox model was assessed by plotting a negative logarithm of the Kaplan Meier Survivor estimate. Hazard ratio in each patient was calculated, and then patients were divided into two groups in the median of hazard ratio. a) A graph assessed the proportionality of hazards in Model 2 divided into high and low hazard. b) A graph assessed the proportionality of hazards in Model 3 in the same way. 10) Urinary phosphate increase by high protein intake might due to the increased uptake, as it is associated with protein rich foods. Can the authors comment? There is a strong correlation between amount of urinary phosphate and protein intake. Thus in our study, fixed amount of protein was served to avoid potential confounds. Discussion: 11) Why is phosphate in serum only increased in Quartile 4? Is this the reason that U-P/24h-CCr is not associated with serum phosphorous in CKD stage 3? In early stage of CKD, serum phosphorus was reported to be kept in normal range due to a increase in phosphate excretion by fibroblast growth factor-23 (FGF-23). But as CKD progresses, FGF-23 cannot enhance urinary phosphate excretion, thus leading to the development of hyperphosphatemia. In this study, we assume serum phosphorus is kept in normal range in CKD stageG3b by increasing U-P/24hCCr. In Quartile 4, because they had lower eGFR than other groups, they had hyper phosphatemia. 12) What is the outccome of an association of fractional phosphorous excretion with CKD progression? The work of Bech 2015 shows no impact on CKD progression by fractional phosphate excretion, which needs to be discussed here. 14) Which information provides U-P/CCr that is described by the established maker, i.e. fractional excretion, TRP and phosphate in serum? As no data of fractional phosphate excretion is provided here, it is difficult to evaluate the better impact of U-P/CCr on CKD progression. Answer to 12) &14) The outcome of an associateon of fractional phosphorus excretion (FeP) with serum phosphorus was shown as follows: Table. Hazard Ratio (HR) and 95% Confidence Interval (CI) of the Composite Outcome of ESKD or 50% Reduction of eGFR Associated with fractional phosphorus excretion Model 1 Model 2 Model 3 HR (95%CI) P Value HR (95%CI) P Value HR (95%CI) P Val ue FeP Quartil e 1 1.00 (referenc e) 1.00 (reference) 1.00 (reference) FeP Quartil e 2 2.82 (130 – 6.77) 0.01 1.57 (0.65 – 4.35) 0.33 1.72 (0.65 – 5.10) 0.2 9 FeP Quartil e 3 7.06 (3.43 – 16.44) < 0.001 2.65 (1.05 – 7.65) 0.03 3.51 (1.23 – 11.4) 0.0 1 FeP Quartil e 4 7.69 (3.63– 18.20) < 0.001 2.46 (0.85 – 7.92) 0.09 2.27 (0.69 – 8.27) 0.1 8 The association between FeP and the outcome was observed in Model 1, but in Model 2 and 3, higher FeP did not showed a higher risk for the outcomes. The work of Bech 2015 shows no impact on CKD progression by FeP. In this study, patients were categorized by FGF23 and FeP. Whereas there was not adjustment for FGF23 in our study. U-P/24h-CCr would be a confounding factor for FGF23. Although U-P/24h-CCr would be a reflecting FGF23, We think that U-P/24h-CCr could be a surrogate marker of FGF23 in clinical practice, because FGF23 cannot be measured in clinical use in Japan. We have changed the following text from p10 line 2 to To compare 24h U-P/CCr with common markers of urinary phosphorus excretion, we examined for the association of fractional phosphate excretion (FeP), We have also added the following text p13 line12 A previous study showed no impact on CKD progression by FeP(). In this study, patients were categorized by value of FGF23 and FeP. In our study, there was not adjustment for FGF23. 24h U-P/CCr would be a confounding factor for FGF23. Even though 24h U-P/CCr would be a reflecting FGF23, We think that U-P/24h-CCr could be a surrogate marker of FGF23 in clinical practice, because FGF23 cannot be measured in clinical use in Japan. Moreover, the association between FeP and the outcome was observed in Model 1, but in Model 2 and 3, higher FeP did not showed a higher risk for the outcomes. We have added the following reference. 21. Bech AP, Bouma-de Krijger A, van Zuilen AD et al. Impact of fractional phosphate excretion on the relation of FGF23 with outcome in CKD patients. J Nephrol 2015 Feb 21. [Epub ahead of print] 13) The alternate model which normalises to GFR showed similar results. Why is U-P/CCr preferred to normalisation to GFR? It is difficult to collected 24-hour urine accurately because of residual urine. The unit of U-P/24h-CCr and U-P/eGFR are presented as follows: (mg/24 hours) / (ml/min/24 hours) and (mg/24 hours) / (ml/min). If collected urine is sampled only 23 hours, the value of U-P/24h-CCr is not change greatly, because U-P and 24h-CCr were collected during the same time. In this case, the unit of U-P/24h-CCr is (mg/23hours) / (ml/min/23hours) But when normalizes eGFR, each urine sample is required to be collected