Impaired Tubular Secretion of Organic Solutes in Advanced Chronic Kidney Disease

Significance Statement

In patients with CKD, the clearance of waste solutes removed by tubular secretion may be altered to an extent that is disproportionate to the reduction in the GFR. However, an average change in the clearance of secreted waste solutes relative to the GFR in CKD has not been reported, possibly because studies performed so far have included few subjects with advanced CKD. The authors found that the secretory clearance of many waste solutes is reduced relative to the GFR in patients with an eGFR<12 ml/min per 1.73 m². As patients approach dialysis, to the extent that secreted solutes contribute to uremic symptoms, reductions in fractional clearances of secreted solutes might cause such symptoms to increase out of proportion to the reduction in GFR.

Visual Abstract

Abstract

Background

The clearance of solutes removed by tubular secretion may be altered out of proportion to the GFR in CKD. Recent studies have described considerable variability in the secretory clearance of waste solutes relative to the GFR in patients with CKD.

Methods

To test the hypothesis that secretory clearance relative to GFR is reduced in patients approaching dialysis, we used metabolomic analysis to identify solutes in simultaneous urine and plasma samples from 16 patients with CKD and an eGFR of 7±2 ml/min per 1.73 m² and 16 control participants. Fractional clearances were calculated as the ratios of urine to plasma levels of each solute relative to those of creatinine and urea in patients with CKD and to those of creatinine in controls.

Results

Metabolomic analysis identified 39 secreted solutes with fractional clearance >3.0 in control participants. Fractional clearance values in patients with CKD were reduced on average to 65%±27% of those in controls. These values were significantly lower for 18 of 39 individual solutes and significantly higher for only one. Assays of the secreted anions phenylacetyl glutamine, p-cresol sulfate, indoxyl sulfate, and hippurate confirmed variable impairment of secretory clearances in advanced CKD. Fractional clearances were markedly reduced for phenylacetylglutamine (4.2±0.6 for controls versus 2.3±0.6 for patients with CKD; P<0.001), p-cresol sulfate (8.6±2.6 for controls versus 4.1±1.5 for patients with CKD; P<0.001), and indoxyl sulfate (23.0±7.3 versus 7.5±2.8; P<0.001) but not for hippurate (10.2±3.8 versus 8.4±2.6; P=0.13).

Conclusions

Secretory clearances for many solutes are reduced more than the GFR in advanced CKD. Impaired secretion of these solutes might contribute to uremic symptoms as patients approach dialysis.

The kidney removes waste solutes from the blood by a combination of glomerular filtration and tubular secretion. The clearance of solutes removed only by filtration remains equal to the GFR when kidney function is impaired. The clearance of solutes removed by secretion may, however, be altered out of proportion to the GFR. Recent studies have described considerable variability in the secretory clearance of waste solutes relative to the GFR in patients with CKD.^1–3 They have further shown that patients with CKD in whom secretory clearance is reduced relative to the GFR may suffer more rapid loss of renal function and greater mortality.^3,4 An average reduction in the secretory clearance of waste solutes relative to GFR as CKD advances has, however, not been detected. This may be because studies reported so far have included few participants with advanced CKD. This study tested the hypothesis that secretory clearances relative to GFR decline as patients approach dialysis. We first measured the clearance in advanced CKD of selected organic anions, which the normal kidney clears rapidly by secretion. Metabolomic analysis was then used to test whether secretion of a broader range of solutes is reduced relative to the GFR in patients approaching dialysis.

Methods

Patient Enrollment

Sixteen patients with advanced CKD characterized by eGFR<12 ml/min per 1.73 m² and 16 control participants without known kidney disease were included in the study. Written informed consent was obtained from all participants. The study was approved by the Stanford Institutional Review Board and was conducted in accordance with the Declaration of Helsinki. Patients with CKD were recruited from renal clinics, and control participants were recruited from among acquaintances and from among patients attending an endocrine clinic to increase the number of patients with diabetes in the control group.

Chemical Analysis

Simultaneous spot urine and plasma samples were obtained from patients with CKD and control participants. Free solute concentrations were measured in plasma ultrafiltrate obtained using Nanosep 30 K Omega separators (Pall, Ann Arbor, MI). Relative concentrations of solutes in the plasma, plasma ultrafiltrate, and urine were assessed by metabolomic analysis. This analysis was performed at Metabolon Inc. using the established platform, which uses liquid chromatography/mass spectrometry to identify solutes from among a chemical library including over 4000 confirmed metabolites.^5,6 Concentrations of selected uremic solutes were also assayed using chemical standards. Urea was measured by a commercial enzymatic assay, and creatinine was measured by HPLC as previously described.⁷ The organic anions phenylacetyl glutamate, hippurate, indoxyl sulfate, and p-cresol sulfate were measured using stable isotope dilution liquid chromatography with tandem mass spectrometry as previously described.⁷

Calculations

Solute clearances relative to the GFR were expressed as fractional clearances. Fractional clearance (FC) values in control participants were estimated by comparing the ratios of urine to plasma levels for each solute with those for creatinine (cre):

$$\text{FC}=\frac{\left[\text{solute}\right]\text{urine}/\left[\text{solute}\right]\text{plasma}}{\left[\text{cre}\right]\text{urine}/\left[\text{cre}\right]\text{plasma}}.$$ (1)

Fractional clearances values in patients with advanced CKD were estimated by comparing the ratios of urine to plasma levels for each solute relative to the mean of the ratios of urine to plasma levels for urea and creatinine:

$$\text{FC}=\frac{\left[\text{solute}\right]\text{urine}/\left[\text{solute}\right]\text{plasma}}{((\left[\text{cre}\right]\text{urine}/\left[\text{cre}\right]\text{plasma})+(\left[\text{urea}\right]\text{urine}/\left[\text{urea}\right]\text{plasma}))/2}.$$ (2)

Creatinine alone was used in calculating fractional clearance values in control participants because the creatinine clearance tends to exceed the GFR so that fractional clearances calculated in this manner would tend to underestimate the true values.⁸ Creatinine together with urea was used in calculating fractional clearances in patients with advanced CKD because the average of the creatinine and urea clearances has been considered to be closer to the GFR in this setting.⁹ The use of creatinine alone for control participants and the combination of creatinine and urea for patients with CKD were adopted to minimize the risk of falsely identifying impaired secretion in advanced CKD as further shown in Supplemental Tables 1–3. Fractional clearances in control participants and patients with CKD were calculated both in terms of the free, unbound solute concentrations measured in the plasma ultrafiltrate and the total concentrations measured in the plasma.

Analysis of metabolomic results was limited to solutes for which peak areas were detected in at least ten of 16 samples of both urine and plasma ultrafiltrate in each of the control participants. This yielded a group of 303 solutes for which fractional clearance values were compared in control participants and participants with CKD as listed in Supplemental Table 4. Among this list, 39 solutes were identified as secreted on the basis of the finding of a fractional clearance >3.0 expressed in terms of the free plasma level as further shown in Supplemental Table 4. Fractional clearances for solutes identified by metabolomic analysis were calculated as described above using the mass spectrometric peak areas to represent solute concentrations in the plasma, plasma ultrafiltrate, and urine.

Fractional clearances of solutes identified by metabolomic analysis in the patients with CKD and control participants were compared using the unpaired t test with the statistical significance assigned after correction for multiple comparisons by the Bonferroni method. Values for the fractional clearances of solutes measured by assays using chemical standards and for other parameters in the patients with CKD and control participants were compared using the unpaired t test. Confidence intervals of the ratios of plasma solute concentrations measured using chemical standards in CKD as compared with control participants were calculated using the Fieller method.¹⁰

Results

The characteristics of the patients with CKD and control participants are summarized in Table 1. Age, weight, and sex were well matched. The prevalence of diabetes was slightly greater among the controls. The plasma creatinine and urea N concentrations averaged 8.0±1.8 and 83±27 mg/dl, respectively, in the patients with CKD as compared with values of 0.9±0.2 and 15±2 mg/dl, respectively, in the control participants. eGFR values calculated by the Chronic Kidney Disease Epidemiology Collaboration equation were 7±2 and 83±15 ml/min per 1.73 m², respectively. Average plasma albumin concentrations were significantly lower in the patients with CKD than in the controls.

Table 1.

Characteristics of participants

Characteristic	CKD	Control
Age, yr	59±14	64±8
Weight, kg	88±23	87±27
Men/women	8/4	8/4
Diabetes	11/5	8/8
Plasma creatinine, mg/dl	8.0±1.8	0.9±0.2
Plasma urea nitrogen, mg/dl	83±27	15±2
eGFR, ml/min per 1.73 m2	7±2	83±15
Albumin, g/dl	3.7±0.4a	4.4±0.5

Results are mean ± SD. eGFR was calculated using the Chronic Kidney Disease Epidemiology Collaboration formula.

^a P=0.05 CKD versus control for albumin. Other chemical values are different by design.

Fractional clearance values obtained with assays using chemical standards are summarized in Table 2. As expected, the fractional clearance of urea was <1.0 in the two groups, indicating tubular reabsorption of urea. In contrast, the fractional clearance of phenylacetylglutamine was >1.0 in both groups, indicating tubular secretion. Phenylacetylglutamine does not exhibit significant protein binding so that fractional clearance values calculated in terms of the total plasma level and free plasma level were nearly the same. The fractional clearance of phenylacetylglutamine in control participants was close to 4.0, indicating that tubular secretion provides a clearance of this organic solute that is close to the renal plasma flow. However the fractional clearance of phenylacetylglutamine was markedly reduced in patients with advanced CKD, indicating that injury reduced secretion out of proportion to the GFR.

Table 2.

Fractional clearances of solutes assayed using chemical standards

Solute	Fractional Clearance				Free Fraction, %
	Control	CKD	CKD/Control	P Value	Control	CKD	P Value
Urea N	0.53±0.16	0.72±0.08	1.4
PAG
Total	3.9±0.8	2.2±0.6	0.57	<0.001	95±17	97±7	0.73
Free	4.2±0.6	2.3±0.6	0.56	<0.001
PCS
Total	0.18±0.04	0.19±0.05	1.04	0.69	2.2±0.4	4.9±1.2	<0.001
Free	8.6±2.6	4.1±1.5	0.47	<0.001
IS
Total	0.53±0.11	0.40±0.14	0.76	0.007	2.4±0.7	5.7±1.7	<0.001
Free	23.0±7.3	7.5±2.8	0.33	<0.001
HIPP
Total	3.4±1.2	4.0±1.0	1.21	0.92	33±2	49±7	<0.001
Free	10.2±3.8	8.4±2.6	0.82	0.13

Results are mean ± SD. The free fraction represents the proportion of solute that is not bound to protein and was calculated as the free level divided by the total level for each solute. CKD/control is the ratio of the average value in the participants with CKD to the average value in the control participants. PAG, phenylacetylglutamine; PCS, p-cresol sulfate; IS, indoxyl sulfate; HIPP, hippurate.

Findings for the tightly bound solutes p-cresol sulfate and indoxyl sulfate provided further evidence of reduced secretion in advanced CKD. When calculated in terms of the total plasma clearance, the fractional clearance of p-cresol sulfate was not reduced in patients with CKD compared with controls. However, the free, unbound concentration of p-cresol sulfate was increased relative to the total concentration in patients with CKD so that the fractional clearance calculated in terms of the free, unbound p-cresol sulfate concentration was markedly reduced. The fractional clearance of indoxyl sulfate was significantly reduced in patients with CKD compared with controls even when calculated in terms of the total plasma level. As with p-cresol sulfate, however, there was a marked increase in the free fraction of indoxyl sulfate so that the fractional clearance expressed in terms of the free, unbound indoxyl sulfate concentration was more severely reduced.

Findings for hippurate differed notably from those for phenylacetylglutamine, p-cresol sulfate, and indoxyl sulfate. Hippurate is partially bound to plasma proteins, and its protein binding was reduced in the patients with CKD like that of p-cresol sulfate and indoxyl sulfate. However, the fractional clearance values for hippurate in the patients with CKD were not significantly different from those in the control participants when expressed in terms of the total plasma hippurate concentration or the free plasma hippurate concentration.

The findings of variable reduction in secretory clearance for solutes assayed using chemical standards prompted a metabolomic analysis of secretory clearance in advanced CKD. Fractional clearance values for the 39 solutes identified as secreted by the normal kidney are summarized in Table 3 and depicted in Figure 1. The fractional clearance values for the 39 secreted solutes in patients with CKD averaged 65%±27% of the values in control participants when calculated in terms of the free concentrations. The values in patients with CKD were significantly lower for 18 of the 39 solutes and significantly higher for only one. Fractional clearance values calculated in terms of the total plasma concentrations were consistent with the values calculated on the basis of the free plasma concentrations as further summarized in Table 3. The extent of protein binding tended to decrease in patients with CKD for solutes that exhibited marked protein binding in control participants. The free, unbound fractions for the seven solutes with a free fraction <0.15 in the control participants were increased by an average two-fold higher in the patients with CKD. Metabolomic analysis identified impaired secretion for five of these seven solutes when clearance was calculated in terms of the free, unbound solute level but not in terms of the total plasma solute level. Metabolomic analysis and assays using chemical standards yielded similar results for solutes examined by both methods as summarized in Supplemental Table 5. In particular, there was no tendency for metabolomic analysis to exaggerate the impairment in the secretory clearance of phenylacetylglutamine, p-cresol sulfate, and indoxyl sulfate in CKD.

Table 3.

Metabolomic analysis of solute secretion

Solute	Fractional Clearance [Free]		Free Fraction		Fractional Clearance [Total]
	Control	CKD/Control	Control	CKD/Control	Control	CKD/Control
1,3,7-Trimethylurate	15.1±5.5	0.76	1.48±0.72	0.6	23.9±9.8	0.76
3-Indoxyl sulfate	14.2±4.8	0.43a	0.03±0.01	2.1a	0.5±0.1	0.85
Pyridoxate	13.1±7.8	0.60
N-(2-furoyl)glycine	11.8±8.1	0.34
Guanidinoacetate	10.1±8.5	0.87	1.16±0.26	0.9	13.9±9.9	0.52
3-Hydroxyhippurate	9.7±6.3	0.93	0.39±0.13	1.2	4.8±3.3	0.86
5-Acetylamino-6-formylamino-3-methyluracil	9.6±5.0	0.64	1.32±0.49	0.7	14.8±6.2	0.37a
4-Methylcatechol sulfate	8.7±4.0	0.29a	0.04±0.01	2.6a	0.4±0.1	0.70
5-Hydroxy-2-methylpyridine sulfate	7.8±5.6	0.28	0.17±0.07	1.8	1.9±0.9	0.31a
Hippurate	7.5±2.7	0.83	0.29±0.07	1.5	2.7±1	1.08
1,7-Dimethylurate	6.4±1.7	0.94	0.29±0.08	1.5a	2.3±0.5	1.20
4-Hydroxyhippurate	6.2±2.5	0.85	0.46±0.14	1.1	3.4±1.4	0.88
N2,N5-diacetylornithine	6.1±3.0	0.43a	0.99±0.31	0.9	7.5±3.6	0.32a
p-Cresol sulfate	5.8±1.2	0.65a	0.04±0.01	1.6a	0.3±0.1	0.96
N-acetyl-2-aminooctanoateb	5.6±4.6	0.46	0.09±0.02	2.2a	0.5±0.4	0.90
Glutamine conjugate of C7H12O2b	5.2±1.9	0.54	0.54±0.16	1.2	3.5±1.1	0.55a
N2,N2-dimethylguanosine	5.2±1.0	0.63a	0.57±0.07	1.2	3.7±0.5	0.70a
3-Methyl catechol sulfate (1)	5.1±1.9	0.59	0.03±0.01	1.7a	0.2±0.1	0.55
2-Isopropylmalate	5.0±4.1	0.88	0.43±0.23	1.4	3.1±2.6	0.79
Acisoga	5.0±3.2	0.84
Vanillylmandelate	4.8±1.0	0.68a	0.74±0.21	1.1	4.4±1.1	0.67a
1-Methylnicotinamide	4.8±1.3	1.58	0.85±0.07	1.0	5.2±1.4	1.47
3-Methoxycatechol sulfate (1)	4.7±2.0	0.50a	0.14±0.04	1.9	0.8±0.2	0.82
1-Methylurate	4.6±1.4	0.78	0.56±0.13	1.2	3.1±1.1	0.84
N-acetylasparagine	4.6±1.6	0.27a	0.8±0.34	0.8	4.3±2.3	0.19a
4-Guanidinobutanoate	4.5±1.7	0.47a	0.85±0.28	1.1	4.9±3	0.36a
Guaiacol sulfate	4.5±1.2	0.49a	0.15±0.04	1.7a	0.8±0.3	0.72
p-Cresol glucuronideb	4.3±1.2	0.52a	0.63±0.13	1.3	3.5±1.4	0.57a
2,6-Dihydroxybenzoic acid	4.1±1.9	0.83	0.16±0.04	1.4	0.8±0.4	1.02
Phenylacetylglutamine	4.0±1.1	0.73	0.69±0.07	1.1	3.4±0.8	0.71a
Hexanoylglutamine	4.0±1.5	0.57a	0.54±0.18	1.3	2.5±0.8	0.66a
1-Methylxanthine	3.9±1.6	0.60
N-acetyltaurine	3.7±0.9	0.49a	0.33±0.15	1.7a	1.7±0.6	0.69
Glutamine conjugate of C6H10O2 (1)b	3.5±0.7	0.50a	0.62±0.19	1.3	2.6±0.8	0.57a
N-acetylglucosamine/N-acetylgalactosamine	3.4±1.0	0.97	0.9±0.17	1.1	3.7±0.9	0.95
4-Hydroxyphenylacetylglutamine	3.4±0.9	0.45a	0.69±0.17	1.1	3±0.9	0.42a
1-Methyl-5-imidazoleacetate	3.3±2.4	0.28a	0.95±0.2	0.9	3.7±2.2	0.24a
Glutamine conjugate of C6H10O2 (2)b	3.3±1.2	0.61a	0.62±0.18	1.2	2.4±1	0.67
3-Aminoazepan-2-one	3.2±0.6	1.3a	0.88±0.14	1.0	3.5±0.7	1.15

Solutes identified as secreted in the metabolomic analysis are listed in order of their average estimated fractional clearance values expressed in terms of the free solute levels in the control participants, denoted fractional clearance [free]. CKD/control is the ratio of the average value in the patients with CKD to the average value in the control participants.

^aP=0.05 for the corresponding value in patients with CKD compared with the value in control participants. The free fraction represents the proportion of solute that is not bound to protein and was calculated as the free level divided by the total level for each solute. Fractional clearance [total] gives the averaged estimated fractional clearance values expressed in terms of the total plasma solute levels in the control participants.

^bA compound whose identity is considered known but has not been confirmed by use of a chemical standard.

Figure 1.

Impaired fractional clearances in CKD of solute identified by metabolomic analysis. Fractional clearance values in patients with CKD are on the y axis and fractional clearance values in control participants are on the x axis for the 39 solutes identified as secreted in control participants. Most of the values fall below the diagonal line of identity, with the values in participants with CKD significantly different from those in control participants for 18 solutes (filled circles) and not significantly different for 21 solutes (open circles).

For the uremic solutes assayed using chemical standards, impaired secretory clearance was accompanied by accumulation in the plasma of patients with CKD as summarized in Table 4. Plasma concentrations of phenylacetylglutamine, free p-cresol sulfate, and free and total indoxyl sulfate for which fractional clearance values were significantly reduced in CKD were elevated out of proportion to concentrations of urea, whereas plasma concentrations of hippurate were not. For other solutes identified as secreted by metabolomic analysis, the pattern of accumulation in the plasma of patients with CKD was less clearly related to their fractional urinary clearance as summarized in Supplemental Table 6. The average concentrations of many of these solutes did not rise as much as the average level of urea in patients with CKD, despite their having high fractional clearance values in control participants and on average, a reduction in their fractional clearance values in patients with CKD.

Table 4.

Plasma solute levels

Solute	Plasma Solute Level
	Control	CKD	CKD/Control
Urea N, mg/dl	15±2	83±27	6 [5 to 7]
PAG, µg/dl
Total	57±43	1335±830a	23 [14 to 40]
Free	50±33	1302±844a	26 [16 to 42]
PCS, µg/dl
Total	394±210	3065±1512a	8 [5 to 12]
Free	9±5	152±98a	18 [11 to 28]
IS, µg/dl
Total	92±41	1516±696a	16 [12 to 23]
Free	2±1	89±52a	38 [24 to 60]
HIPP, µg/dl
Total	154±120	1071±816a	7 [4 to 13]
Free	49±36	537±446a	11 [6 to 20]

Plasma total and free levels of solutes measured using chemical standards. The 95% confidence intervals for ratios are in brackets. PAG, phenylacetylglutamine; PCS, p-cresol sulfate; IS, indoxyl sulfate; HIPP, hippurate.

^aP<0.001 CKD versus control. Urea N is considered different by design.

Discussion

Recent studies have emphasized the importance of tubular secretion in the kidney’s clearance of waste solutes.^11–13 This study found that the tubular secretion of such solutes is impaired in advanced CKD. It extends the results of prior studies of tubular secretion in CKD.^1–3 Those studies documented variability in secretory clearance as kidney function declines but no overall tendency toward reduction in secretory clearance relative to GFR. They included, however, very few participants whose eGFR values were as low as those described here.

A notable finding of this study was that impaired secretion of protein-bound solutes in CKD was more apparent when clearances were expressed in terms of free, unbound solute concentrations than in terms of total solute concentrations. The combination of rapidly reversible protein binding and tubular secretion normally keeps the free concentrations of such solutes to which body cells are exposed very low. Clearances expressed in terms of the free solute concentrations can exceed the renal plasma flow, as reflected by fractional clearance values greater than five for free p-cresol sulfate, indoxyl sulfate, and hippurate in this study. As reported in previous studies, however, solute binding to proteins can be impaired in renal failure so that the free concentrations of bound solutes increase relative to the total concentrations in the plasma.^1,14,15 Impaired binding may result from competition among solutes for binding sites or from a chemical alteration in albumin, which is presumed to be the main binding protein. Whatever its cause, impaired protein binding results in a lesser increase in total solute concentrations than in free solute concentrations. Calculation of clearance values in terms of total solute concentrations thus tends to obscure an impairment in tubular secretion, which limits the kidney’s capacity to maintain low free concentrations of waste solutes.

Although reduced secretory clearance of endogenous solutes in advanced CKD has not previously been described, reduced secretory clearance of exogenous solutes has been observed. The most extensively studied secreted solute is para-aminohippuric acid (PAH). In normal humans, PAH is extracted almost completely by secretion from blood flowing through the peritubular capillaries. The clearance of PAH has, therefore, long been used to estimate the renal plasma flow. Early studies showed that PAH clearance generally falls in proportion to the GFR in patients with CKD but may be disproportionally reduced in patients with advanced disease.^16,17 More recent studies have demonstrated disproportionate reduction in the secretory clearance of several drugs in patients with advanced CKD.^18,19 Impaired secretion in these drugs can be considered analogous to the impaired secretion of endogenous solutes observed in this study.

The mechanisms by which the secretory clearance of some solutes is reduced out of proportion to GFR in advanced CKD remain to be elucidated. Structural changes may contribute. Fenestrated capillary endothelial cells provide close apposition of capillary plasma to the basal surfaces of proximal tubular cells in the normal kidney. In CKD, this interface is disrupted by the associated phenomena of interstitial expansion and capillary rarefaction.^20,21 These changes may impede access to the tubular cells of both secreted solutes and metabolites needed for energy production. Reduced expression of transport proteins could further impair secretory clearance. Expression of most transporters has not been examined in CKD, but reduced fractional clearance of hippurate has been associated with reduced expression of the organic anion transporter 1 (OAT1) in rats subjected to partial renal ablation.^22,23

A further potential contributor to decreased fractional clearance of secreted solutes in advanced CKD is their increasing plasma concentrations. Clearance by secretion, unlike clearance by glomerular filtration, is saturable. As their plasma concentrations rise, secreted solutes compete with themselves and with each other for transport by a limited number of secretory pathways.^24–27 Complex structural characteristics determine preferential transport by different pathways, and drugs excreted by the same pathways may inhibit the excretion of endogenous solutes in patients with CKD.^28–30 Known examples include diuretic drugs, which were used by many of our patients with CKD.³⁰ Competition for secretion may result in the clearance of some solutes being impaired more than others. Previous studies have shown that hippurate and indoxyl sulfate are both transported by OAT1.^22,28 We found that secretion of hippurate was not significantly impaired in our patients with advanced CKD, whereas the fractional clearance of indoxyl sulfate was markedly reduced. This could reflect preferential transport by OAT1 of hippurate over indoxyl sulfate as concentrations of both compounds rise. Both competition with other solutes and transport by different pathways could account for the variable reduction in the secretory clearances of individual solutes that we observed in advanced CKD. The same factors presumably account for the variable effect of CKD on the secretory clearances of different drugs.^18,19 Drugs that impair creatinine secretion could also cause us to overestimate the fractional clearance of secreted solutes in patients with CKD.

An interesting finding of our metabolomic study was that plasma concentrations did not rise in proportion to the reduction in clearance for some secreted solutes. We presume that this reflects reduced production of these solutes in CKD and/or the availability of other routes for their clearance. A reduction in fractional clearance will cause plasma concentrations to rise out of proportion to the fall in GFR for solutes whose clearance is accomplished exclusively by the kidney and whose production is maintained in CKD. When combined with results of prior studies, our findings suggest that tubular secretion of waste solutes is impaired only in patients with advanced CKD. We speculate that in such patients, competition for secretory transport can drive plasma solute concentrations much higher if the GFR falls only a little farther or if drugs secreted by the same pathways are administered. If secreted solutes contribute to uremic symptoms, such an increase in their concentrations would cause symptoms to increase out of proportion to any fall in the GFR.

Our study has important limitations. Metabolomic analysis by untargeted mass spectrometry can identify hundreds of solutes in individual samples. The metabolomic platform we used, however, is by no means complete. Solute levels are, moreover, represented only by peak areas, and absolute chemical concentrations are not obtained. The peak areas obtained at a given solute concentration can be strongly affected by other compounds in the solution being analyzed and may not be linearly associated with absolute solute concentrations in the same solution. Metabolomic analysis can thus provide only estimates of the ratios of urine to plasma levels used to calculate fractional clearances in this study. We confirmed the accuracy of the metabolomic analysis with assays using chemical standards for selected solutes. These solutes were, however, all organic anions with mass between 180 and 250 D. Secreted solutes with different chemical characteristics may behave differently as CKD advances. The fractional clearance of some such solutes may indeed be increased. Our metabolomic analysis identified one such potential solute, 3-aminoazepan-2-one, for which standard databases provide references only to chemical syntheses and potential toxicity. Finally, analysis of spot urine samples obtained near the time of plasma sampling for calculation of fractional clearance values will fail to identify diurnal changes in solute production.

In summary, we found that secretory clearances for many solutes are reduced more than the GFR in advanced CKD. Further studies will be required to determine if impaired secretion contributes to uremic illness as patients approach dialysis.

Disclosures

T.W. Meyer reports research funding from Outset Medical; honoraria from Baxter; and scientific advisor or membership via the JASN and Kidney International editorial boards. T.W. Meyer and T.L. Sirich have served as consultants for Baxter. All remaining authors have nothing to disclose.

Funding

This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Ruth L. Kirschstein National Research Service Award F32 DK111166 (to R.D. Mair), grant R01 DK101674 (to T.W. Meyer), and grant R01 DK118426 (to T.L. Sirich). S. Lee was supported by an American Society of Nephrology Ben J. Lipps Fellowship.

Supplemental Material

This article contains the following supplemental material online at https://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ASN.2021030336/-/DCSupplemental.

Supplemental Table 1. Fractional clearances of solutes assayed using chemical standards calculated using U/P_creatinine for both control participants and participants with CKD.

Supplemental Table 2. Fractional clearances of solutes assayed using chemical standards calculated using U/P_urea for both control participants and participants with CKD.

Supplemental Table 3. Ratios of average fractional clearance values in participants with CKD as compared with control participants using different denominators to estimate the GFR.

Supplemental Table 4. Solutes for which metabolomic analysis provided a comparison of fractional clearance values in control participants and patients with advanced CKD.

Supplemental Table 5. Comparison of fractional clearance values obtained by metabolomic analysis and by assays using chemical standards.

Supplemental Table 6. Plasma concentrations in patients with advanced CKD relative to control participants of solutes identified as secreted in control participants.