Functional Reserve of the Kidney

Abstract

An exploration of the normal limits of physiologic responses and how these responses are lost when the kidney is injured rarely occurs in clinical practice. However, the differences between “resting” and “stressed” responses identify an adaptive reactiveness that is diminished before baseline function is impaired. This functional reserve is important in the evaluation of prognosis and progression of kidney disease. Here, we discuss stress tests that examine protein-induced hyperfiltration, proximal tubular secretion, urea-selective concentration defects, and acid retention. We discuss diseases in which these tests have been used to diagnose subclinical injury. The study and follow-up of abnormal functional reserve may add considerable understanding to the natural history of CKD.

Introduction

Stress tests that challenge the physiologic responses of the kidney identify a functional reserve that is seldom explored in clinical practice; yet, these tests deserve special attention. Blunted or lost differences between a normal “resting” and “stressed” response uncover subclinical injury that may be of considerable importance as a risk or prognostic factor in the natural history of CKD.

In this review, we extend the concept of kidney functional reserve beyond its common usage of protein-induced hyperfiltration. Tubular function does not follow a simple correlation with GFR (¹,²), and tubular dysfunction resulting in retention of protein-bound uremic toxins constitutes independent mortality and cardiovascular risk factors (³). Therefore, we include tests that challenge the proximal tubular secretion mediated by organic transporters, more recent tests that have identified urea-selective concentration impairment, and tests that uncover normocarbonatemic (normal serum bicarbonate level) acid retention. In addition, we review the conditions in which these tests have been associated with subclinical injury.

Protein-Induced Glomerular Hyperfiltration

The initial studies of kidney functional reserve in humans were done by Bosch et al. (⁴), who tested the hyperfiltration induced by a protein meal. Subsequent studies expanded these observations and, despite numerous investigations examining circulating levels and/or blocking the effects of glucagon, glucagon-inulin ratio, vasopressin, PGs, dopamine, angiotensin, nitric oxide, somatostatin, growth hormone, and IGF, the physiopathology of protein-induced hyperfiltration remains incompletely understood (reviewed in ^5–8). Figure 1 shows an incomplete list of systemic hormonal changes that drive a reduction of the tubuloglomerular feedback restraint of the GFR. This reduction results from the generation of nitric oxide in the macula densa, mediated by upregulation of neuronal nitric oxide synthase β and phosphorylation of serine 1417 (⁹). Both afferent and efferent glomerular arterioles are dilated. The renal vascular resistance is decreased and the renal blood flow and GFR are increased, with stable filtration fraction (Figure 1).

Figure 1.: Protein-induced hyperfiltration results from the complex interplay of systemic hormonal changes (glucagon, vasopressin, IGF, somatostatin, growth factor, increased urea synthesis by the liver) that induce in the kidney functional changes that drive a reduction in the tubuloglomerular feedback (TGF) restraint of the GFR, with similar afferent and efferent glomerular arteriolar vasodilation. The reduction on the sodium concentration in the urine reaching the macula densa, which inhibits the TGF, results from the interplay of increased proximal reabsorption of sodium, increased urinary urea concentration, and robust activity of the sodium-potassium-chloride cotransporter 2 (NKCC2) in the thick ascending limb of the loop of Henle. Prostaglandins, kinins, and angiotensin participate in the functional changes in a manner incompletely understood. The results are a reduction in renal vascular resistance, and comparable increments in renal blood flow and GFR that maintain a stable filtration fraction.

Table 1 shows the studies of protein-induced hyperfiltration done by several groups (⁴,¹⁰,^13–28,³⁰,³¹), including ourselves (¹¹,¹²,²⁹), in health and disease using oral protein loads and amino acid infusions with and without dopamine. The use of creatinine clearance in some studies results in inconsistent results because high-protein meals increase the tubular secretion of creatinine (³²). Additional variability is caused by the lack of uniformity in the protein load administered; the prior protein content of the diet; the status of hydration (overhydration reduces the protein-induced hyperfiltration and most studies have been done with water loading); and the ingestion of medications, such as nonsteroidal anti-inflammatory agents, that modify the GFR response. Nevertheless, the variability of the normal response in studies that used radioisotope or inulin clearances to estimate the GFR is not unacceptably high. Although intraperson variability has not been explored, the interperson coefficient of variation of stimulated GFR in healthy subjects ranged from 0.10 to 0.17 in most studies (¹⁰,¹¹,¹⁴,¹⁸,²⁴). Only two studies (¹⁵,²³) reported a coefficient of variation in controls of >0.20.

Table 1. - Studies of protein-induced hyperfiltration

Clinical Condition (reference)	N	Method	Results	Comments
Healthy adults (4)	5	80 g cooked red meat	Stimulated creatinine clearance; 160±7.7 ml/min	Reduced response in four of six patients with kidney disease
Healthy adults (10)	10	3.5 g/kg cooked red meat	GFR max of 132±6 ml/min	Increased kidney blood flow with reduced vascular resistance
Normal adults (11)	27	Large or mild oral protein load	GFR max of 141±7.69 ml/min (large protein load)	Large variability; filtration fraction was increased at peak GFR levels
School children (12)	260	Oral 200–300 g lean hamburger	Stimulated creatinine clearance of 143.1±2.24 ml/min	Stimulated creatinine clearance negatively correlated with baseline creatinine clearance
Pregnancy (13)	29	Oral protein 0.3–1.2 g/kg	Mean increase in creatinine clearance of 63 ml/min	Higher glomerular functional reserve in third trimester
Pregnancy (14)	10	300 g red meat	Increase of GFR from 106±5 to 119±4 ml/min	GFR increment in gestation same as nonpregnant individuals
Pregnancy (15)	21	Amino acid infusion	GFR increase of 10%–18%	GFR increment same as in nonpregnant individuals
Pregnancy (16)	8	Oral protein 80 g	Creatinine clearance increased; 97±10.1 ml/min	50% reduction in pregnant patients with hypertension
Kidney donors (17)	Meta-analysis of 1547 (1425 pre- and poststudies)	Dopamine, infusions, amino acid infusions, oral protein	Predonation GFR increment; 8%–23%	Postdonation GFR increment reduced; 1%–25%; glomerular functional reserve reduced in donors who were obese, hypertensive, and elderly
Kidney donors (18)	125 donors	Amino acid infusion, dopamine	Baseline GFR of 112±18 ml/min; stimulated GFR of 138±22 ml/min	Predonation glomerular functional reserve did not add value to the prediction of postdonation CKD
Kidney transplant recipients (19)	12	Amino acid infusion	GFR increase of 17–28 ml/min	Unaffected by CyA treatment
Hemolytic uremic syndrome a (20)	26	Oral protein	Glomerular functional reserve reduced; 36%
CKD stages 1–4 (21)	100	Amino acid infusions	Progressive reduction in glomerular functional reserve with the severity of CKD
DM1 (22)	15	Amino acid infusions	Mean GFR increase of 6%	GFR increase of 13% in normal individuals
DM1 (23)	19	Amino acid infusion	No glomerular functional reserve	GFR increase of 26% in normal individuals
DM1 (24)	51	Dopamine infusion	Absent glomerular functional reserve	Normal glomerular functional reserve of 21±8 ml/min
DM2 (25)	18	Amino acid infusion	Glomerular functional reserve (White patients, 21.9±45.7 ml/min; Asian African patients, −2.5±28.2 ml/min)	African American patients with diabetes lose glomerular functional reserve earlier than White patients
AKI (26)	86	Oral protein	36% mean reduction of glomerular functional reserve estimated by creatinine clearance	Patients recovered from AKI had reduced glomerular functional reserve
Hypertensive risk (genetic) (27)	39	Amino acid infusions	Glomerular functional reserve of 2%	Glomerular functional reserve in controls; 31%
Systemic sclerosis (28)	21	Amino acid infusions	GFR increase of 2%	GFR increase in controls; 35%
Postacute streptococcal GN (29) b	35	Oral protein	Creatinine clearance increase of 18%	Creatinine clearance increase in controls; 48%
Postacute streptococcal GN (30) b	36	Oral protein	Creatinine clearance increase of 18%	Creatinine clearance increase in controls; 41%
Posthemolytic uremic syndrome (31) a	33	Oral protein	50% reduction

Glomerular functional reserve glomerular in normal and clinical conditions. GFR determined by inulin clearance or radioisotopic methods. Glomerular functional reserve refers to the increase over baseline GFR. GFR and clearances shown corrected for 1.73 m². CyA, cyclosporine A; DM1, diabetes mellitus type 1; DM2, diabetes mellitus type 2.

^aPatients recovered from hemolytic uremic syndrome.

^bPatients recovered from poststreptococcal GN.

Hyperfiltration is stronger with a protein meal than with intravenous amino acids (³³), and the rise in GFR is absent with branched-chain amino acids (³⁴). The increment in GFR typically ranges from 20%–30%. Notably, expressing stimulated GFR as a percentage change would give lower values if the baseline GFR is as high as it may be when the prior diet has a high protein content or in conditions such as early diabetes. While filtration fraction is generally unaltered in protein-induced hyperfitration, increments in filtration fraction may be found exceptionally at high levels of hyperfiltration induced with a meal of cooked red meat (¹¹). Given all of these considerations, we agree with Palsson and Waikar (⁶) on the need to establish uniform guidelines for the evaluation of protein-induced hyperfiltration.

Table 1 shows that impaired response to protein loading is found in patients with CKD stage 1 and CKD stage 2 (²¹), those with diabetes type 1 and 2 (^22–25), those with a genetic risk of hypertension (²⁷), hypertensive kidney donors (¹⁷), those with systemic sclerosis (²⁸), and in patients after recovery from AKI (²⁶). A similar impaired response was found after recovery in patients with acute GN (²⁹,³⁰) and in those with Shiga toxin–induced hemolytic uremic syndrome (³¹). In these patients, the clinical significance of the reduced functional reserve deserves to be investigated. It should be noted that patients recovered from acute poststreptococcal GN have a two-hit risk of CKD in association with obesity, diabetes, and hypertension (³⁵). Special mention is deserved of the study by Livi et al. (²⁸), who studied the glomerular functional reserve in 28 patients with systemic sclerosis, without urinary abnormalities, and with normal kidney function, and did a follow-up study 5 years later. Nineteen patients who originally showed a reduced functional reserve had a more rapid deterioration of kidney function and higher incidence of albuminuria and hypertension. Clearly, follow-up studies are necessary to confirm that impaired protein-induced hyperfiltration represents a risk marker of CKD progression.

Recently, image technologies have been used to evaluate the GFR increment. Reductions of 22%–25% in resistive and pulsatility indexes determined by ultrasound correlated with GFR changes (³⁶,³⁷). The evaluation of kidney vasodilation with positron-emission tomography has been suggested as a means to test kidney functional reserve (³⁸). These findings require validation in larger studies.

Stress Tests of Proximal Tubule Organic Transporters

The removal of protein-bound solutes by the kidney depends on a system of organic anion transporters (OATs) and organic cation transporters (OCTs) that facilitate the basolateral uptake and luminal secretion of compounds with negative charge (anionic), positive charge (cationic), and with both negative and positive charge (zwitterions). Both OCTs and OATs are transporters of complex overlapping specificities reviewed competently elsewhere (^39–41). OATs have overlapping specificities with furosemide, methotrexate, nonsteroidal anti-inflammatory drugs, thiazide diuretics, angiotensin II antagonists, β-lactam antibiotics, and antiviral drugs. The OCT system is responsible for the secretion of cationic substrates. In humans, the most important OCT is OCT2, which is responsible for creatinine secretion and has overlapping specificities with many drugs (cimetidine, trimethoprim, metformin, atenolol, albuterol, amiloride, triamterene, metoclopramide, procainamide, and others) and uremic toxins that impair OCT function (guanidine, polyamines) (⁴²).

A large number of solutes, including indoxyl sulfate, p-cresyl sulfate, homocysteine, hippuric acid, pyridoxic acid, kynurenic acid, and trimethyl uric acid, are accumulated in uremia, circulate with protein binding ranging from 51%–96% (⁴³,⁴⁴), and depend mostly on tubular secretion for their disposal. These compounds have a very limited clearance by dialysis membranes because of the size of the protein-compound complex, but their protein binding is competitive and reversible. In a proof-of-concept study, we demonstrated that infusion of ibuprofen, administered as a competitive binder, increased more than two times the hemodialysis removal of protein-bound uremic toxins. Clearly, potential ibuprofen toxicity prevents its use as therapeutic strategy (⁴⁵).

Acute ischemia of the kidney causes downregulation of mRNA and expression of tubular transporters and accumulation of the protein-bound solutes (⁴⁶). These findings may predict the outcome of AKI (⁴⁷). The remnant kidney and parenchymal kidney disease result in reduced expression of OATs (⁴⁸,⁴⁹). Studies from our group (^50–52) evaluated the proximal tubule functional reserve determining the increase in tubular secretion of creatinine after an exogenous creatinine challenge (Table 2). Because substrate-stimulated tubular transport is only found in the adult kidney (⁵⁴), the increase in creatinine secretion is likely the result of directing available transport activity to meet the acute demands imposed by the substrate load. It is expected that the recruitment of transport capacity toward one ligand would result in a reduced secretion of other ligands that share the same transport system. We use the term “tubular functional reserve” for the increased excretion of the loaded substrate because an impaired response is the best indication we have of dysfunction or reduction in the promiscuous tubular transporters.

Table 2. - Studies of stimulation of organic transporters in proximal tubule

Clinical Condition (reference)	N	Method	Results	Comment
Healthy adults, K donors, and patients with CKD (50)	12 normal individuals, 7 kidney donors, 8 patients with CKD	Oral protein 80 g	Tubular secretion of creatinine: normal patients, 114.4±12.65 nmol/kg per min; K donors, 76.6±13.34 nmol/kg per min; patients with CKD, 5.53±19.66 nmol/kg per min a	Patients with CKD did not increase tubular secretion of creatinine
Healthy adults, kidney donors, and K transp (51)	14 normal patients, 7 kidney donors, 11 transplant recipients	IV creatinine (88 μmol/kg)	Tubular secretion of creatinine: normal patients, 180±60 nmol/kg per min; K donors, 155±54 nmol/kg per min; K transp, 86±35 nmol/kg per min a	Tubular secretion of creatinine is a better index of reduced functioning kidney mass than GFR
Sickle cell disease (52)	16 patients with sickle cell disease, 20 normal patients	IV creatinine (88 μmol/kg)	Tubular secretion of creatinine: patients with sickle cell disease, 123±52 nmol/kg per min; control patients, 179±50 nmol/kg per min a	Hyperfiltering patients with sickle cell disease had reduced tubular reserve function
CKD (53)	84 patients (grades 1, 2, and 3 fibrosis)	IV furosemide (1 mg/kg or 1.5 mg/kg)	Urine output correlated with fibrosis score and furosemide excretion mass	Furosemide excretion mass may reflect tubular functional reserve

Functional reserve of proximal tubule organic transporters. Cationic transporters evaluated by the creatinine secretion stimulated by oral or intravenous creatinine. Anionic transporters evaluated by the tubular stress response to intravenous furosemide. K donors, kidney donors; K transp, kidney transplant recipients; IV, intravenous.

^aTubular secretion of creatinine 90 minutes after the administration of an intravenous bolus of 88.4 μmol creatinine/kg body wt.

In our studies, the tubular secretion of creatinine was estimated as the difference between the urinary excretion of creatinine (urinary creatinine concentration×urine volume) and the filtered creatinine (serum creatinine concentration×GFR). When healthy individuals were given a meal of 80 g of animal protein, their tubular creatinine secretion increased by several folds to values of 114.4±12.7 nmol of creatinine/min per kg in the following 90 minutes (Table 2). Patients with kidney disease and kidney donors had a reduced tubular creatinine secretion. These studies were expanded to include the use of intravenous bolus of 88.4 μmol of creatinine/kg body wt (⁵¹). This dose of creatinine resulted in serum creatinine levels of 500–700 μmol/L, shown by us and others (⁵⁵,⁵⁶) to result in the highest creatinine clearance/inulin clearance ratios in healthy individuals, indicating a maximum tubular transport for creatinine. We found there was a six-fold increase in tubular creatinine secretion, without significant changes in GFR, in the 30 minutes that followed the creatinine load (Figure 2A). The independence of GFR and the response to the creatinine stress test is further demonstrated in Figure 2B, which includes data from patients with a single kidney (kidney donors) and data from 16 patients with homozygous sickle cell disease (⁵²). The patients with sickle cell disease are known to have glomerular hyperfiltration in adolescence and deterioration of kidney function usually by the age of 40 (⁵⁷,⁵⁸). Distal nephron alterations (incomplete tubular acidosis, decreased concentrated ability) are frequent in these patients, but proximal tubular function, as evaluated by p-aminohippuric acid and uric acid secretion, has been found to be increased. As shown in Figure 2B, young hyperfiltering patients with sickle cell disease showed a 40%–50% reduction in the stimulated tubular creatinine secretion, indicating the OCT2-dependent tubular functional reserve was exhausted. These findings are likely the consequence of repeated episodes of tubulointerstitial ischemia, as demonstrated to occur in transgenic sickle cell mice (⁵⁹).

Figure 2.: Kidney tubular functional reserve is independent of GFR changes. Data show tubular functional response 30 minutes after an intravenous bolus load of 88.4 μmol creatinine/kg body wt that achieved maximal stimulation of creatinine secretion (see text). (A) GFR, tubular secretion of creatinine (TScr), and TScr per milliliter of GFR (TScr/GFR) in healthy individuals. The lack of significant GFR changes is in contrast with the six-fold increase in TScr. (B) Comparison of the response of TScr in normal individuals, patients with sickle cell disease (SCD), kidney donors (KDonors), and transplanted patients (KTx; 4–164 months after transplantation) of comparable age. Data demonstrate the independence of stimulated TScr from GFR (inulin or iodine-125–iothalamate clearance). Drugs that interfere with creatinine secretion were suspended before the studies. Figure drawn from data in Rodriguez-Iturbe et al. (⁵¹) and Herrera et al. (⁵⁰). GFR=mean±SEM; TScr=mean±95% confidence interval (normal individuals versus those with SCD and KTx; P<0.01).

Another way to test the response of tubular organic transporters is the recently described furosemide stress test. Chawla et al. (⁶⁰) used a furosemide stress test to predict the progression in severity of AKI. Furosemide diuretic action depends on its active secretion by OAT in the proximal tubule for its subsequent inhibition of the sodium-potassium-chloride cotransporter 2 in the thick ascending limb of the loop of Henle. The test consisted of measuring diuresis after a furosemide dose of 1.5 mg/kg or 1.0 mg/kg, depending on whether the patient had or had not, respectively, previously received loop diuretics (⁶¹). A urine output of <200 ml in the following 2 hours predicted the progression to AKI stage 3 with a sensitivity of 87% and a specificity of 84% (⁶⁰,⁶²). Using the same furosemide dose, we studied associations between the diuretic response and the extent of interstitial fibrosis in 84 kidney biopsy specimens from patients with nephrotic syndrome and kidney graft dysfunction (⁵³). The patients with the highest degree of fibrosis showed lower urine output in the first hour (P=0.02), and there was a negative linear correlation between the furosemide excreted mass and the degree of fibrosis (r=−0.25, P=0.02). We speculate that exploring the furosemide excreted mass in healthy individuals could give valuable information on the normal function of the OAT system.

The studies in Table 2 show that testing the tubular secretion of creatinine reveals changes in proximal tubular function that are independent of GFR (Figure 2). Furthermore, the abnormal results in patients with CKD (⁵⁰), transplanted kidneys (⁵⁰,⁵¹), and patients with sickle cell disease (⁵²) who have a preserved GFR indicate that stimulating tubular creatinine secretion provides a more sensitive test of impaired kidney function than eGFR. The potential usefulness of this test in the subclinical detection of tubulointerstitial injury deserves to be investigated.

Studies of Prolonged Water Deprivation and Urea Urine/Plasma Concentration Ratio

The urinary osmolarity in humans may vary by more than 20-fold (50–1200 mOsm/kg water) in response to extremes of suppression or stimulation of the secretion of the antidiuretic hormone (arginine vasopressin; AVP) and its interaction with V2 receptors (V2Rs). Water reabsorption depends on two AVP-V2R–related effects: the opening of water-permeable aquaporin channels in the principal cells of the connecting and collecting ducts, and the stimulation of interstitial hypertonicity that increases toward the tip of the medulla. The hyperosmolarity of the interstitium is driven by the active sodium reabsorption driven by sodium-potassium-chloride cotransporter 2 in the water-impermeable thick ascending limb of Henle, and by the building up of the interstitial concentration of urea (⁶³,⁶⁴).

Prolonged Water Deprivation

Studies of prolonged water deprivation have been used to evaluate the urinary concentration in patients with autosomal dominant polycystic kidney disease (ADPKD) (⁶⁵). After 14 hours of water deprivation, patients with ADPKD showed impaired maximal urinary concentration (758 versus 915 mOsmol/kg in controls; P<0.001) in association with higher levels of plasma osmolality and vasopressin levels, indicating an appropriate hypothalamic response to an insufficient capacity to concentrate urine. These results suggested a mechanism for ADPKD progression that is accelerated by the binding of vasopressin to the V2R with production of cAMP, causing cyst enlargement (⁶⁶). These studies of stressed urinary concentration gave support to the use of tolvaptan to slow the progression of ADPKD (⁶⁷).

Urea Urine/Plasma Concentration Ratio

Zittema et al. (⁶⁸) found that when patients with ADPKD were given 2 µg of desmopressin intramuscularly, they had lower urine urea and osmolarity than patients with IgA nephropathy matched for kidney function. These findings likely resulted from the medullary location of cysts in patients with ADPKD. Bankir and Bichet (⁶⁹) analyzed data from the study of Zittema et al. (⁶⁸) and noted the urea urine/plasma concentration ratio was two-fold lower in patients with ADPKD than in the controls with IgA nephropathy (42.4 versus 84.3 mmol/L), whereas the urine/plasma osmolar ratio was only slightly lower in those with ADPKD (2.77 versus 3.58 mOsm/kg water), which was evidence of a “urea-selective” concentrating defect in ADPKD.

These findings underline the importance of evaluating the urinary/plasma solute concentration, in addition to maximal urine osmolarity, when testing the response to AVP. Solutes like urea do not have tightly regulated plasma levels, and a greater fraction of the urea in the interstitial medulla returns to circulation when there are alterations in the countercurrent exchange mechanism, as is the case in ADPKD with medullary cyst formation. More recently, Heida et al. (⁷⁰) reported that the urea urine/plasma concentration ratio in spot urine and plasma samples obtained in the early morning after 12-hour dehydration correlated with the maximal urinary concentration capacity. In addition, they found that the odds ratio of rapidly progressive disease (eGFR reduction >3 ml/min per 1.73 m²) was 1.35 (95% confidence interval, 1.19 to 1.52; P<0.001) for every 10 U decrease in urine/plasma urea concentration ratio, and established that a urea-selective concentration defect represented an additional risk marker of progression in the disease.

Evaluation of Subclinical Acid Retention

The responses of the kidney to maintain acid-base homeostasis include the modulation of the excretion of hydrogen ions (H⁺; titratable acidity), bicarbonate reabsorption, the generation of ammonia, and excretion of ammonium (⁷¹). Ammonium chloride and bicarbonate administration are used to evaluate the different types of tubular acidosis (⁷²,⁷³).

Recent studies have focused on the detection of subclinical acid retention (⁷⁴). This is important because acid retention in CKD starts much earlier than when it is overtly manifested by reduction in serum bicarbonate, which is usually found in CKD stage 4 or later. Metabolic acidosis with normal serum bicarbonate has been evaluated by ammonium-induced urinary citrate excretion and by the estimation of the expected increase in serum bicarbonate concentration after oral bicarbonate administration.

The potential usefulness of studying urinary citrate excretion to evaluate acid retention was demonstrated by Goraya et al. (⁷⁵) who showed that H⁺ retention was higher and urinary citrate was lower in patients with CKD stage 2 than in patients with CKD stage 1 without over metabolic acidosis, and that a base-producing diet reduced H⁺ retention and increased urinary citrate excretion. Acid retention was defined as “unaccounted hydrogencarbonate ion (HCO₃⁻),” as determined by the difference between the expected and the observed increase in plasma bicarbonate after 2 hours of the administration of an oral bolus of sodium bicarbonate (NaHCO₃; 0.5 mg/lean body wt), assuming a bicarbonate distribution space of 50% body weight, i.e., H⁺ retention=([retained HCO₃/0.5×body wt]−observed increase in plasma HCO₃)×0.5 body wt.

Goraya et al. (⁷⁵) estimated that, in patients with CKD stage 2, a “cutoff” value of urinary citrate excretion (U_citrateV) level of 230 mg had a positive predictive value of 91% and a negative predictive value of 71% to predict H⁺ retention.

Urinary Citrate Excretion Induced by Ammonium Chloride Administration

Urinary citrate excretion is controlled primarily by proximal tubular reabsorption (⁷⁶) and, as noted by Gianella et al. (⁷⁷), represents a “defense mechanism” against acid gain that offers a more sensitive measure of the acid retention than changes in the “defended parameter” of plasma bicarbonate. In patients who were stone formers, a retrospective analysis of urinary citrate/creatinine ratio was 187 (interquartile range, 125–277) mmol/mol in patients with creatinine clearance ≥90 ml/min, and this was progressively reduced to 40 (interquartile range, 23–102) mmol/mol in patients with creatinine clearance <30 ml/min. Importantly, reduced serum bicarbonate was evident when the creatinine clearance was <30 ml/min, whereas the reduction in the urine citrate/creatinine ratio was present at earlier stages of CKD.

The increase in net acid excretion (net acid excretion=[titratable acid+ammonium]−[citrate+bicarbonate]) measured before and 4 hours after the oral administration of 50 mEq of ammonium chloride was used to estimate acid retention that was correlated (P=0.04) with the urine citrate/creatinine ratio, but not with eGFR nor with baseline serum bicarbonate levels. In addition, in spot morning urine samples, the urine citrate/creatinine ratio was progressively reduced with the severity of CKD, starting with patients with CKD stage 1. In contrast, mean serum bicarbonate levels were steady until eGFR was reduced to <30 ml/min per 1.73 m². The variability was large: acid retention ranged from 35 to 60 mEq when the urine citrate/creatinine ratio ranged between 200–300 mmol/mol. However, the simplicity of this test for the detection of subclinical acid retention makes it particularly attractive and deserves to be evaluated with prior dietary control and in association with net acid excretion for the potential diagnosis of acidosis with normal serum bicarbonate.

Bicarbonate Treatment in Subclinical Acidosis

The usefulness of bicarbonate treatment in subclinical acidosis was demonstrated by Goraya et al. (⁷⁸) who randomized patients with CKD who were nondiabetic and had a plasma total carbon dioxide >24 mM/L to receive 0.5 mEq/kg body wt daily of NaHCO₃, the same dose of sodium chloride, or usual care (n=40 each group). H⁺ retention was estimated as the difference between the expected and measured increase in serum bicarbonate after an acute oral bicarbonate load, as described earlier. In a 10-year follow-up, it was found that the eGFR was better preserved in the NaHCO₃ group that showed no increase in acid retention during follow-up. Worse preservation of eGFR in the sodium chloride and usual care groups was associated with higher 10-year versus baseline acid retention and lower urinary citrate excretion. These studies show that identification of patients with CKD stage 2 and subclinical acidosis offers the opportunity to improve their long-term prognosis by the administration of oral bicarbonate.

Conclusions

Table 3 summarizes the stress tests that explore the kidney functional reserve discussed in this review. Stress tests are a valuable approach to establish the physiologic limits of nephron function. Further research is required to establish the practical benefits of some of these tests that are cumbersome and time consuming. Nevertheless, there are clinical conditions in which these tests deserve to be used more extensively than they presently are. Associations between impaired stress responses and serum or urine biomarkers of kidney damage need to be investigated. There are limited options for estimating CKD prognosis, and the study of the stress tests of glomerular and tubular function may represent a useful avenue to better define risk factors, subclinical injury, and natural history of kidney disease progression.

Table 3. - Stress tests of kidney function

Stress Test	Response Tested	Normal Response	Clinical Usefulness
Oral protein, IV amino acids ()	GFR increment	10%–30% increase in GFR	Single nephron hyperfiltration
Creatinine load (50,52)	Proximal tubule organic cation transport	Max TScr: oral protein, 114.4±12.7 nmol/kg per min; IV creatinine TScr, 180±60 nmol/kg per min	Association/correlation with retention of protein-bound solutes
Water restriction-AVP–induced urea urine/plasma concentration ratio (69,70)	Urinary concentration, urea countercurrent exchange	Urea U/P ratio ≥80, at copeptin levels of 11.9 (IQR, 7.1–28.3) pmol/L	ADPKD, interstitial nephritis
Oral NH4Cl (50 mEq) and urinary citrate/creatinine ratio (77)	H+ ion retention with normal serum bicarbonate	Urine citrate/creatinine ratio of 187 (IQR, 125–277) mmol/mol a	Treatment of subclinical acidosis
Oral bicarbonate (78,79)	H+ ion retention with normal serum bicarbonate	3±14 mmol H+ ion retention b	Treatment of subclinical acidosis

IV, intravenous; max, maximum; TScr, tubular secretion of creatinine; AVP, arginine vasopressin; U/P, urine/plasma concentration ratio; IQR, interquartile range; ADPKD, autosomal dominant polycystic kidney disease; NH₄Cl, ammonium chloride; H⁺, hydrogen ion.

^aData obtained in 24 urine collections. Urine citrate/creatinine concentration ratio is reduced progressively with increasing CKD severity (see text).

^bValue listed as normal corresponds to the value found in patients with eGFR ≥90 ml/min per 1.73 m² (CKD stage 1) in Goraya et al. (⁷⁹).

Disclosures

M. Madero reports serving on advisory boards of Abbvie, AstraZeneca, and Bayer; receiving research funding from Abbvie, AstraZeneca, Bayer, and Boehringer; serving as a scientific advisor or member of American Journal of Kidney Disease, International Society of Nephrology, and Kidney Disease Improving Global Outcomes Executive Committee; speakers bureau for AstraZeneca; and receiving honoraria from AstraZeneca, Baxter, and Fresenius Medical Center. All remaining authors have nothing to disclose.

Funding

This work was supported by Fundacion Gonzalo Rio Arronte grant 93131700.