Abstract
The optimal approach to managing acid-base balance is less well defined for patients receiving hemodialysis than for those receiving peritoneal dialysis. Interventional studies in hemodialysis have been limited and inconsistent in their findings, whereas more compelling data are available from interventional studies in peritoneal dialysis. Both high and low serum bicarbonate levels associate with an increased risk of mortality in patients receiving hemodialysis, but high values are a marker for poor nutrition and comorbidity and are often highly variable from month to month. Measurement of pH would likely provide useful additional data. Concern has arisen regarding high-bicarbonate dialysate and dialysis-induced alkalemia, but whether these truly cause harm remains to be determined. The available evidence is insufficient for determining the optimal target for therapy at this time.
Metabolic acidosis was one of the earliest recognized complications of uremia.1 Several of the initial reports of dialysis therapy demonstrated its correction by diffusion of bicarbonate from dialysis fluid to uremic blood.2–4 In the decades since, significant technologic improvements have occurred, but the basic concept for acid-base regulation remains the same. Thus it is somewhat surprising that controversy persists about the proper prescription of bicarbonate dialysis and even the appropriate target for therapy.
This was aptly illustrated in the last several years. In 2012 a large dialysis provider faced a spate of lawsuits stemming from a memo addressing acid-base status and the use of its acetate-containing dialysates.5,6 The memo recommended changes to dialysis prescriptions on the basis of an association of high serum bicarbonate with mortality. For many nephrologists in the United States, this case created confusion about bicarbonate dialysis and added medico-legal significance to the hemodialysis prescription. It is useful to consider why such uncertainty still exists.
Rationale for Treatment of Metabolic Acidosis in ESRD
Patients with ESRD appear to be in net acid-base balance.7 Thrice-weekly hemodialysis creates a cycle of rapid correction of acidosis that causes transient alkalemia of varying degree, followed by acid retention during the ensuing 44–68 hours. Patients treated with peritoneal dialysis experience nearly continuous buffer repletion. In neither group does net acid retention occur.7,8 However, sequelae of their residual acidosis were identified, including altered protein, muscle, and bone metabolism.9–12 Correction of metabolic acidosis in patients with ESRD improved insulin sensitivity and reduced whole-body and skeletal muscle protein breakdown.13–17 It also enhanced parathyroid gland sensitivity to calcium and improved bone turnover in both high- and low-turnover states.18,19 However, the findings of interventional studies in patients receiving hemodialysis were somewhat inconsistent, and were limited by small sample sizes and lack of a control group in some.20–25
The most compelling data came from two trials in patients treated with continuous ambulatory peritoneal dialysis. In a randomized, double-blind, placebo-controlled trial of 60 patients, those assigned to oral sodium bicarbonate had improved nutritional status and experienced fewer hospitalizations after 1 year.26 Correction of acidosis in another single-blind randomized trial of 200 patients receiving peritoneal dialysis increased lean body mass and reduced hospitalizations.27 The latter intervention targeted a serum bicarbonate of 30 mEq/L using dialysate with 40 mmol/L lactate (compared with 35 mmol/L in the control group), calcium carbonate, plus oral sodium bicarbonate as needed. Notably, the intervention groups in both studies achieved mean serum bicarbonate levels of approximately 28 mEq/L or higher, which not only was well tolerated but which resulted in improved outcomes. Peritoneal dialysis solutions containing 40 mmol/L lactate are now standard.
Serum Bicarbonate Levels in Patients Receiving Hemodialysis
As these findings were reported, predialysis serum bicarbonate levels among patients receiving hemodialysis increased. Among 12,099 patients receiving hemodialysis in the United States in 1987–88, the mean serum bicarbonate was 19.2 mEq/L28; in contrast, the mean serum bicarbonate was 21.4 mEq/L in 4515 United States patients receiving hemodialysis in the Dialysis Outcomes and Practice Patterns Study (DOPPS) from 1996–2001,29 and was 21.9 mEq/L among 56,385 patients who received hemodialysis from 2001–03.30 More striking is that the percentage of patients with serum bicarbonate <21 mEq/L declined from 75% in the 1987–88 cohort to 35% in the 2001–03 cohort.28,30
Several factors explain this trend. The switch from acetate to bicarbonate as the main source of dialysate buffer replacement improved correction of acidosis.31–33 The use of higher blood flows and high-efficiency membranes would be expected to increase bicarbonate delivery, but may not have had an effect.34 Nephrologists may have targeted low serum bicarbonate values more aggressively, especially after an initial observational study that associated levels of 20–22.5 mEq/L with the lowest risk of death.28 In 2000 the National Kidney Foundation Kidney Disease Outcomes Quality Initiative (KDOQI) recommended maintaining predialysis serum bicarbonate at ≥22 mEq/L,35 which would have further spurred treatment of low bicarbonate levels. Secular trends also could have played a role, such as the increasing age of patients initiating dialysis in the United States. Older Americans have higher serum bicarbonate levels than younger Americans, partly due to lower animal protein intake and hence lower dietary acid load.36 Older age is associated with higher serum bicarbonate in patients with ESRD as well.37–39 Therefore, it is conceivable that the increasing age of incident dialysis patients has contributed to increasing serum bicarbonate levels.
Despite these changes, a substantial proportion of patients do not meet the KDOQI target. From 2001 to 2006 at one large dialysis provider, 40% of 110,951 patients receiving hemodialysis had a time-averaged predialysis serum bicarbonate <22 mEq/L.39 Only 5% had serum bicarbonate ≥27 mEq/L.39 Nevertheless, changes in practice due to recent concerns about high serum and dialysate bicarbonate levels could increase the proportion of patients with low predialysis serum bicarbonate levels.
Pre- and Postdialysis Bicarbonate Levels and Mortality in Patients Receiving Hemodialysis
Both high and low serum bicarbonate levels have been associated with an increased risk of mortality in observational studies of patients receiving hemodialysis (total carbon dioxide [CO2], generally 1–1.5 mEq/L higher than the serum bicarbonate concentration, is actually what is measured in all epidemiologic studies unless otherwise noted).28–30 The association of low bicarbonate with mortality is magnified after adjustment for markers of nutrition, comorbidity, and inflammation, whereas that of high bicarbonate is markedly attenuated.30,39 This suggests that high predialysis serum bicarbonate is a marker for patients with poor nutrition, ongoing inflammation, and a high burden of comorbidity. Indeed, serum bicarbonate correlates inversely with normalized protein catabolic rate, serum albumin, and serum phosphate in patients with ESRD.30,37,38 Put simply, sicker patients eat less protein, resulting in less acid generation. Thus the association is unlikely to represent a causal relationship.
Some have expressed concern about the use of high dialysate bicarbonate or total alkali (the sum of bicarbonate plus an organic anion, such as acetate or citrate) concentrations on the basis of these findings. In general, patients being dialyzed against a high-bicarbonate bath do have higher predialysis bicarbonate levels.24,25,39 However, there are clearly other important determinants of predialysis serum bicarbonate (Table 1). The United States and Japan were at opposite ends of the spectrum in terms of dialysate bicarbonate levels from 1996 to 2001 (highest and lowest, respectively) but had the same mean serum bicarbonate—21.4 mEq/L, or the lowest of seven nations surveyed.29 A number of factors account for this finding. Whereas lower blood flows in Japan would reduce net bicarbonate gain, longer treatment times and smaller average body size compared with the average United States patient receiving hemodialysis would increase net bicarbonate transfer into a smaller bicarbonate space.40,41 Daily acid production is likely lower for the average Japanese patient due to the lower dietary acid load of the Japanese diet, further reducing the amount of buffer repletion needed to achieve a given bicarbonate level.
Determinants of pre- and posthemodialysis serum bicarbonate
Differences in endogenous acid production, largely determined by dietary intake, account for some of the interindividual variability in serum bicarbonate levels.30,34,37,38 Interdialytic weight gain lowers the bicarbonate level by expanding the bicarbonate space.34,42–44 Serum bicarbonate is approximately 1 mEq/L lower after the long interdialytic interval than the short interval.29,45,46 Medication use is important as well; sevelamer hydrochloride reduces serum bicarbonate, whereas phosphate binders containing alkali precursors do the opposite.47–51
Net alkali gain during dialysis and its volume of distribution determine the postdialysis serum bicarbonate (Table 1). This sets the starting point for the linear decline that most patients experience between treatments.45 The change in serum bicarbonate during hemodialysis correlates with the dialysate–serum bicarbonate gradient (Figure 1A) and relates inversely to the predialysis level (Figure 1B), so patients with the highest levels at the initiation of treatment experience modest increases (Figure 2).52–54 In a stable chronic hemodialysis patient, net alkali gain during dialysis should equal the total net acid production in the interdialytic period. When measured, this quantity has been quite variable among patients.53,55 In some individuals, the change in bicarbonate may be blunted due to organic anion generation.34 Because this variability has not been quantified, its significance remains unclear. Dialysis dose and ultrafiltration rate, by affecting bicarbonate delivery, are also important (the latter reduces net bicarbonate gain). Finally, in a minority of patients the bicarbonate level does not decrease appreciably between treatments.45
Change in serum bicarbonate during hemodialysis correlates directly with dialysate-serum gradient and inversely with predialysis concentration. Intradialytic change in serum bicarbonate versus (A) dialysate–serum bicarbonate gradient and (B) predialysis serum bicarbonate level among chronic hemodialysis patients dialyzed against 35 mEq/L (15 patients, Uribarri et al.); 36 mEq/L (70 patients, Sepandj et al.); and 25, 30, or 35 mEq/L (53 patients, Noh et al.) bicarbonate dialysate. Data from Uribarri et al.,53 Sepandj et al.,54 and Noh et al.52 Intradialytic increase in serum bicarbonate per 1 mEq/L higher predialysis serum bicarbonate = −0.55 mEq/L (95% confidence interval [95% CI], −0.68 to −0.43); for individual studies, the slope was −0.54 mEq/L (95% CI, −0.81 to −0.28) for Uribarri et al., −0.55 mEq/L (95% CI, −0.69 to −0.41) for Sepandj et al., and −0.43 mEq/L (95% CI, −0.66 to −0.20) for Noh et al. P for interaction by study = 0.63.
Theoretical bicarbonate profiles in hemodialysis. Profiles of serum bicarbonate changes for patients receiving hemodialysis with midweek predialysis serum bicarbonate of 20 and 28 mEq/L, respectively. Profiles assume dialysate bicarbonate of 35 mEq/L and intradialytic increases on the basis of point estimates derived from linear regression of data from Uribarri et al.53 and Sepandj et al.54 (predicted intradialytic increase 7.7 mEq/L [95% confidence interval, 7.2 to 8.3] and 2.8 mEq/L [95% confidence interval, 2.1 to 3.5] for predialysis serum bicarbonate of 20 and 28 mEq/L, respectively).
A recent study of 15,132 Japanese patients receiving hemodialysis provided important new data by examining blood gas measurements at the beginning and end of hemodialysis.56 Although the aforementioned differences between the United States and Japan may limit generalizability, these data provide insights into hemodialysis-induced acid-base changes in a large cohort of patients. Surprisingly, neither high pH nor serum bicarbonate measured at the end of dialysis associated with mortality. Rather, patients in the highest pH quartile before dialysis (pH ≥7.40) had the greatest mortality after multivariable adjustment. As expected, predialysis bicarbonate was higher for these patients than the rest of the cohort, but the partial pressure of carbon dioxide (pCO2) was lower, suggesting a degree of respiratory alkalosis. Indeed, in a small cohort of Italian patients receiving hemodialysis, predialysis pH ≥7.40 most often indicated respiratory alkalosis, sometimes with a concomitant metabolic disorder.46 Predialysis serum bicarbonate was also a poor predictor of predialysis pH.46 These data show that the association of high predialysis serum bicarbonate with mortality is not mediated by alkalemia due to excessive alkali delivery.
Rather, they support the hypothesis that underlying comorbidity explains the mortality risk observed with predialysis alkalemia and high serum bicarbonate. The mean pCO2 rose during dialysis in the group with predialysis alkalemia.56 This suggests they are sicker and less able to appropriately increase minute ventilation in the face of increased CO2 generation with bicarbonate dialysis.57 Higher pCO2 levels could explain the nonsignificant trend toward increased mortality among patients with the lowest pH (<7.40) after dialysis. Indeed, higher normalized protein catabolic rate, a marker for better nutritional status, was unexpectedly associated with a greater likelihood of postdialysis alkalemia. This is surprising because it also associated, as expected, with lower serum bicarbonate postdialysis. The association with postdialysis pH can then only be explained by changes in CO2 excretion during dialysis. The implication is that patients with better nutritional status are more likely to manifest an appropriate respiratory response during hemodialysis.
Finally, the Japanese data do not support the hypothesis that patients with high predialysis serum bicarbonate are most likely to experience severe alkalemia during a hemodialysis treatment. The difference in pH between the highest and lowest bicarbonate quartiles was 0.10 pH units before dialysis but only 0.03 pH units after dialysis.56 That the pH difference shrank during hemodialysis suggests that patients with high predialysis serum bicarbonate experience the least fluctuation in pH, consistent with the data on intradialytic change in serum bicarbonate described above. Therefore, the totality of available data indicates that high predialysis bicarbonate levels are not a useful biomarker for guiding changes in the bicarbonate dialysis prescription.
Prescription of Bicarbonate Hemodialysis
There are still other reasons for concern over the use of high-bicarbonate dialysis. In a DOPPS report of 17,031 patients from 11 countries, higher dialysate bicarbonate concentrations were associated with an increased risk of all-cause mortality.58 Several mechanisms could explain these findings, none of which has been examined beyond small studies. Hypokalemia, reduction in ionized calcium, and prolongation of the QT interval due to rapid or excessive alkalinization could predispose to cardiac arrhythmia59–61; this would be linked specifically to cardiac arrest or sudden death during or in the hours immediately after a dialysis treatment. The DOPPS results appeared to be driven by an increase in infection-related mortality, which is not compatible with this hypothesis. High-bicarbonate dialysis may have a mild vasodilatory effect.62 In patients with intradialytic hypertension, this could be beneficial, but in others symptomatic hypotension may result.63,64 Over time, intradialytic hypotension may induce ischemic cardiac and cerebral injury, predisposing to cardiac arrhythmias, sudden death, and cognitive decline.65,66 That said, hemodialysis-induced left ventricular systolic dysfunction was not associated with change in pH or bicarbonate in a study of 105 patients receiving dialysis.67
Intradialytic hypoxemia could exacerbate ischemic injury and has been associated with mortality.68 Conceivably, hypoxemia could accompany the intradialytic hypercarbia noted above. Studies in patients with COPD have been inconclusive and data from patients with severe comorbidity are warranted.69–71 Finally, recurrent extracellular alkalinization could promote vascular calcification.72,73 This has never been thoroughly investigated in humans. The association with mortality in DOPPS was strongest among patients with longer dialysis vintage, and both vascular calcification and the sequelae of chronic ischemia would manifest after prolonged exposure to high-bicarbonate dialysate.
Although the DOPPS investigators found no difference in associations between the United States and other countries, it is important to recognize the drastic differences in practice patterns worldwide. The United States, which has higher ESRD-adjusted mortality rates than Europe or Japan,74,75 accounted for the majority of patients dialyzed with baths containing ≥38 mEq/L bicarbonate. In contrast, nearly 70% of German patients were dialyzed against bath bicarbonate ≤32 mEq/L. Japan is even more extreme (and was excluded from the recent DOPPS report); dialysate bicarbonate ≤30 mEq/L was prescribed for 83% of patients. To some extent, dialysate bicarbonate prescription could simply track with other country-level factors that associate with mortality.
The Anion in the Acid Concentrate
Bicarbonate was the buffer used when hemodialysis was first introduced.76 It was replaced by acetate due to the latter’s ease of use and the belief that the capacity for its metabolism far exceeded delivery.77 As technology progressed to enable modern proportioning systems and it became apparent that some patients experienced toxicity due to insufficient rates of acetate metabolism,78 bicarbonate again became the primary buffer. When admixed with an acid concentrate containing acetic acid, the final dialysate contains acetate concentrations too low to produce blood levels that clearly associate with toxicity.78 Modest elevations in blood acetate may be seen, but this can be observed even with bicarbonate dialysis absent an organic anion.79
The effect on acid-base balance is not clear. One would expect that the higher the acetate concentration in the dialysate, the more rapidly serum bicarbonate will rise. As a result, the dialysate–plasma bicarbonate gradient will decline more rapidly, so the net effect on serum bicarbonate should diminish with longer treatment time. Consequently, this likely affects early acid-base changes during a dialysis treatment but overall acid-base balance less so. Data from a subset of 7572 patients from the Japanese cohort support this hypothesis. Both pre- and postdialysis serum bicarbonate, as well as the intradialytic change, were lower among patients dialyzed against a 30 mEq/L bicarbonate bath versus 35 mEq/L, despite equal dialysate alkali content (6 mEq/L acetate and 1 mEq/L citrate were added to each, respectively).56 The dialysate bicarbonate is clearly the most important consideration.
Solutions are now available that use citric acid rather than acetic acid as the acidifying agent.80 These may improve dialysis efficiency, which has been attributed to reduced dialyzer clotting due to citrate’s anticoagulant effect. Greater buffer delivery and better correction of acidosis would be expected, but this has not been a consistent finding.81–85
Acid-Base Therapy in ESRD
Current guidelines mostly recommend maintaining serum bicarbonate ≥22 mEq/L in patients with CKD (the European Best Practice Guidelines suggest maintaining midweek predialysis serum bicarbonate at 20–22 mEq/L) on the basis of observational studies and small interventional studies.35,86–89 Thus the available evidence base provides only general guidance for clinicians, leaving room for substantial variability in clinical practice. For example, in some dialysis units the dialysis bicarbonate prescription is individualized, whereas in others a uniform prescription is used for all or nearly all patients.58 There are no trial data to support one practice over the other.
Several lines of evidence point to benefit from full correction of metabolic acidosis. In the two randomized trials of patients receiving peritoneal dialysis discussed earlier, the mean serum bicarbonate in the intervention groups reached levels that have been associated with increased mortality in patients receiving hemodialysis, yet they experienced improved outcomes.26,27 In the study that supplemented high lactate dialysate with oral alkali, the mean pH and serum bicarbonate at 1 year were 7.44 and 27.2 mEq/L, respectively, and this was the nadir bicarbonate level during the intervention.27 Mild alkalemia was associated with improved nitrogen balance in another study of patients receiving peritoneal dialysis, indicating decreased protein breakdown from full correction of acidosis.90 As associations of serum bicarbonate levels with mortality do not differ between peritoneal dialysis and hemodialysis patients, this suggests that high bicarbonate levels at the end of the interdialytic period are not harmful per se.39
However, observational studies do not associate high predialysis serum bicarbonate with improved outcomes, and we should be cautious about extrapolating data from patients receiving peritoneal dialysis to the hemodialysis population. The former provides a relatively continuous supply of buffer and a closer approximation of physiologic acid-base regulation than hemodialysis. Whereas the serum bicarbonate in a patient receiving peritoneal dialysis approximates steady-state, the predialysis serum bicarbonate in a patient receiving hemodialysis represents the peak of acid retention during the interdialytic period (Figure 2). The time-averaged serum bicarbonate during the interdialytic interval—which can be approximated by the difference between the postdialysis and predialysis bicarbonate level45—might provide a better comparison and treatment guide, although still not representative of a true steady-state.
Serum bicarbonate is not an optimal marker, despite associations with mortality, due to confounding by nutrition and other factors. Predialysis pH would better define a patient’s acid-base status, which could enable clinicians to tailor the prescription to an individual patient. Even a serum bicarbonate <20 mEq/L may be an unreliable predictor of acidemia,46 further underscoring the utility of measuring pH to guide management. Although associations of pH with mortality may be confounded by differences in respiratory status due to comorbidity, pH would still provide useful information for prescribing bicarbonate dialysis. Ideally, predialysis pH should be checked before changing the dialysate bicarbonate concentration or prescribing oral alkali, especially in patients with comorbidities such as congestive heart failure, chronic obstructive pulmonary disease, and cirrhosis.
The current evidence base does not sufficiently define the risks and benefits of aggressively treating metabolic acidosis with high-bicarbonate dialysate or oral alkali. Evaluation of a patient with predialysis serum bicarbonate outside the desired range should therefore focus on modifiable factors that accompany high or low bicarbonate levels in patients receiving dialysis. For high serum bicarbonate, this includes an assessment of nutritional status and a search for sources of inflammation. If the level is persistently high (>28 mEq/L), it seems reasonable to lower the dialysate bicarbonate to 30–32 mEq/L with close monitoring of subsequent serum bicarbonate levels. The evaluation of patients with low bicarbonate levels should focus on ensuring adequate dialysis and limiting interdialytic weight gain. Substituting plant protein for animal protein may be helpful,91 and in select patients, increasing fruit and vegetable intake. For patients with persistently low levels (<19 mEq/L) despite these interventions, base delivery can be increased either by raising the dialysate bicarbonate prescription or by administering oral alkali. The latter has often been frowned upon because of concerns about sodium loading and compounding dialysis patients’ already excessive pill burden.92 Therefore, changing dialysate bicarbonate is the most expeditious maneuver.
Even a modest number of pills, however, might increase serum bicarbonate to a level that a clinician considers acceptable. Four 650 mg tablets of sodium bicarbonate on nondialysis days would completely neutralize the average daily acid production of 28 mEq/d measured by Uribarri et al.53 Even half that amount could have a meaningful effect on acid-base status. Although the data from this small study may not be representative of the wider hemodialysis population, they suggest that net acid production for many patients is lower than what others have estimated.31 Of course, the patients who require such an intervention—those with a persistently low serum bicarbonate despite adequate dialysis—are likely to have above-average acid-generation rates (Figure 3), but a trial of oral sodium bicarbonate could be attempted and then discontinued if ineffective. Such an intervention, if tested, could permit lowering the dialysate bicarbonate and might help avoid both interdialytic acidosis and postdialysis alkalosis.
Daily acid-generation rate correlates inversely with predialysis serum bicarbonate level in patients receiving hemodialysis. All patients with daily acid-generation rates above the mean (28 mEq/L) had predialysis serum bicarbonate <24 mEq/L. Data from Uribarri et al. (n=15).53
In clinical practice, it is not entirely clear when to intervene, as variability within an individual patient may hamper clinical decision-making. We noted substantial month-to-month variability in a local cohort of patients receiving hemodialysis.37 A single abnormal serum bicarbonate (outside a normal range of 22–26 mEq/L for the purposes of this study) was as likely to return to normal the following month as to remain high or low. A low 3-month predialysis mean was more reproducible, but a high 3-month mean was twice as likely to be followed by a normal mean serum bicarbonate during the following 3 months as to remain high. Viewing results in the context of a patient’s prior variability may provide additional guidance, but this requires further study. This confounds not only the decision to begin an intervention, but also how to determine its effectiveness, and further suggests clinicians should be reticent about treating high predialysis serum bicarbonate values per se.
Some variability could be due to measurement error. Several reports identified spuriously low bicarbonate levels in samples drawn at dialysis units and shipped to central laboratories for measurement.93,94 This discrepancy was attributed to differences in assays, delays in sample processing, and loss of CO2 from the sample during shipment.93–95 However, the main culprit may have been the time samples were left uncapped in large commercial laboratories, which reduces the measured bicarbonate approximately 2.5 mEq/L per hour of exposure.96 Practices in the dialysis unit also matter. Partially-filled vacutainer tubes and samples left unspun for prolonged periods both spuriously lower the reported value; significantly underfilled tubes cause clinically meaningful reductions.97–99
Finally, a recent paper proposed a novel solution for minimizing intradialytic alkalosis while still adequately treating acidosis.100 The underlying premise is that expansion of the plasma bicarbonate pool in the early part of a dialysis treatment generates a gradient that is sufficient to replete intracellular buffers. Further alkali delivery causes alkalosis but does not have an important effect on buffer stores. Hence, high dialysate bicarbonate is useful initially but then detrimental, and the dialysate bicarbonate should be progressively lowered throughout the remainder of the dialysis treatment. Further work is needed to determine the utility of this maneuver as well as the specific dialysate modeling approach, but this is an intriguing proposal.
Conclusions
The optimal approach to managing acid-base balance is less well defined for patients receiving hemodialysis than for peritoneal dialysis. Adequately powered studies should test whether dialysis-induced alkalemia causes harm and whether benefit accrues from full correction of acidosis. The utility of using pH and the interdialytic averaged serum bicarbonate (postdialysis bicarbonate − predialysis bicarbonate)45 to guide treatment should also be examined. Optimal treatment may require knowledge of both pre- and postdialysis acid-base parameters, with individualized treatment decisions that weigh the risks of predialysis acidosis, rapid intradialytic alkalinization, and postdialysis alkalosis while accounting for a patient’s comorbidities. Further research would be needed to determine if the time and resources required of such an approach led to improved patient outcomes.
Disclosures
M.K.A. has consulted for Tricida, Inc., South San Francisco, CA.
Acknowledgments
This work was supported by K23 DK099438 from the National Institutes of Health (NIH). Its contents are solely the responsibility of the author and do not necessarily represent the official views of the NIH.
Footnotes
Published online ahead of print. Publication date available at www.jasn.org.
- Copyright © 2017 by the American Society of Nephrology