Abstract
Introduction
Low serum bicarbonate at a single point in time is associated with accelerated kidney decline in patients with chronic kidney disease (CKD). We modeled how changes in serum bicarbonate over time affect incidence of adverse kidney outcomes.
Methods
We analyzed data from Optum’s deidentified Integrated Claims-Clinical data set of US patients (2007–2019) with ≥1 year of prior medical record data, CKD stages G3 to G5, and metabolic acidosis (i.e., index serum bicarbonate 12 to <22 mmol/l). The primary predictor of interest was the change in serum bicarbonate, evaluated at each postindex outpatient serum bicarbonate test as a time-dependent continuous variable. The primary outcome was a composite of either a ≥40% decline in estimated glomerular filtration rate (eGFR) from index or evidence of dialysis or transplantation, evaluated using Cox proportional hazards models.
Results
A total of 24,384 patients were included in the cohort with median follow-up of 3.7 years. A within-patient increase in serum bicarbonate over time was associated with a lower risk of the composite kidney outcome. The unadjusted hazard ratio (HR) per 1-mmol/l increase in serum bicarbonate was 0.911 (95% confidence interval [CI]: 0.905–0.917; P < 0.001). After adjustment for baseline eGFR and serum bicarbonate, the time-adjusted effect of baseline eGFR and other covariates, the HR per 1-mmol/l increase in serum bicarbonate was largely unchanged (0.916 [95% CI: 0.910–0.922; P < 0.001]).
Conclusion
In a real-world population of US patients with CKD and metabolic acidosis, a within-patient increase in serum bicarbonate over time independent of changes in eGFR, was associated with a lower risk of CKD progression.
Keywords: chronic kidney disease, CKD progression, eGFR decline, metabolic acidosis, serum bicarbonate
Graphical abstract
As kidney function declines, patients lose the ability to excrete the daily dietary acid load and develop metabolic consequences. Metabolic acidosis is a common complication in advanced CKD and has been associated with worsening kidney function, early mortality, cardiovascular events, adverse bone and muscle outcomes, and impaired immune response.1, 2, 3
Metabolic acidosis in CKD is caused by diminished net acid excretion by the kidneys.4,5 The deleterious effects of metabolic acidosis on kidney function appears to be mediated by multiple factors, including endothelin-1, activation of the renin-angiotensin aldosterone system,6, 7, 8 and complement activation.9
Previously, several observational studies have examined the association between lower serum bicarbonate and progression of CKD. These studies have found conflicting results, with some demonstrating an increased risk of CKD progression (kidney failure or ≥40% reduction in eGFR),10,11 and some finding no association after adjustment for baseline eGFR.12 Many of these studies were small or from single centers. Recently, a large cohort study of over 51,000 patients with CKD found that a lower serum bicarbonate value at baseline was associated with a decline in eGFR of 40% or greater, or kidney failure.13
Notably, these studies only examined the association of serum bicarbonate at a single point in time, and only 1 study examined the association of repeated bicarbonate measurements over time. This study found a 7% reduction in renal events (initiation of renal replacement therapy (RRT), halving of eGFR, or a 25 ml/min per 1.73 m2 decline of eGFR from baseline), and was done in patients followed-up with as part of the Chronic Renal Insufficiency Cohort, which followed patients enrolled at 7 clinical centers across the US.14 We hypothesized that the duration of time a patient is exposed to retained acid (metabolic acidosis) and an upward or downward trajectory of retained acid is likely to affect the extent of adverse effect of metabolic acidosis on the kidney. Serum bicarbonate can fluctuate because of a variety of factors, including laboratory measurements, variability, volume status, kidney function, concomitant medications, and dietary intake.15,16 Use of multiple bicarbonate measurements over time may, therefore, be a better indicator of risk compared to a single value.
Therefore, we conducted a retrospective cohort study of over 24,000 patients with CKD (eGFR <60 ml/min per 1.73 m2) and metabolic acidosis (serum bicarbonate 12 to <22 mmol/l) with longitudinal serum bicarbonate data to determine the association between changes in serum bicarbonate from baseline and progression of kidney disease.
Methods
Study Design and Patient Population
We conducted an observational, retrospective cohort study by extracting data from the Optum’s deidentified Integrated Claims-Clinical data set of US patients (2007–2019) for individuals with at least 1 year of activity in electronic health record (EHR) data, ≥3 eGFR results <60 ml/min per 1.73 m2 on separate days, and ≥3 serum bicarbonate results, at least one of which was between 12 and 29 mmol/l. The Optum data set is a longitudinal clinical repository including 103 million patients from large health care provider organizations in all 50 states in the United States and Puerto Rico, including patients with all types of insurance as well as uninsured patients.17 Extracted data elements utilized for this analysis were derived from inpatient and outpatient EHRs and administrative systems, including information on outpatient laboratory results, coded diagnoses, procedures from all settings of care, and provider notes extracted by natural language processing. Laboratory tests performed during inpatient admissions or hospital emergency department visits were excluded in the analyses because these may be confounded by an acute concurrent illness or medical procedures. Data cleaning by investigators included exclusion of patients whose data came from a healthcare organization that did not report hospital activity, patients without either diagnosis codes or encounter records, and patients with reported death dates before 2007.
Patients were selected from the data extract at the first outpatient eGFR result <60 ml/min per 1.73 m2 (baseline eGFR) preceded by at least 1 year of patient activity in the EHR data. The closest outpatient serum bicarbonate value within ±180 days of baseline eGFR, that allowed for 1 year of prior data, established the baseline serum bicarbonate value and the study index date. Cohort inclusion further required baseline serum bicarbonate between 12 and <22 mmol/l and no preindex evidence of RRT, defined as renal transplantation (by diagnosis code, procedure code or diagnosis related group) or dialysis (by diagnosis code, procedure code, or outpatient eGFR <10 ml/min per 1.73 m2). eGFR was calculated using the CKD epidemiology collaboration equation.18
Variables
Change in serum bicarbonate from baseline (the postindex value minus the baseline value) was calculated for each postindex outpatient serum bicarbonate test before the outcome event, death, or censoring. Age was evaluated based on year of birth, which was normalized to 1930 for all persons born in 1930 and earlier for privacy reasons. Key comorbidities were established by diagnosis code (atrial fibrillation, coronary artery disease, diabetes, heart failure, hypertension, peripheral vascular disease, and stroke) in the preindex year. Baseline laboratory values were evaluated using the closest outpatient value occurring before the index date during the preindex year (serum calcium, serum albumin, hemoglobin, serum potassium, and serum phosphorous) or using all available preindex data (urine albumin-creatinine ratio). Missing data was present in only laboratory values and was not imputed. Variable sources, definitions, and data cleaning, including validity parameters applied to laboratory values are specified in Supplementary Table S1.
Outcomes
The primary outcome was a composite of either kidney failure (defined as evidence of dialysis or transplantation) or a ≥40% decline in eGFR from baseline (RRT40)13; components were evaluated separately as secondary outcomes. An eGFR decline of ≥40% was identified at the first of 2 consecutive outpatient eGFR tests, each representing a ≥40% decline from baseline. Month and year of death were identified using linked Social Security data before data deidentification.
Last engagement date was established using data points that indicated a hands-on interaction with the patient for determining end dates of the following available EHR data: an observation (e.g., blood pressure or weight), procedure record, hospital or emergency department discharge, laboratory sample collection, or a change in medication with rationale. Patients were followed in the analysis until the outcome, death, or the final such last engagement date.
Statistical Analysis
Patient characteristics are reported using percentages, mean and standard deviation, or median and intraquartile range. Unadjusted outcomes were evaluated as event rates per 100 patient-years, with years measured between the index date and the last date of EHR data available for the patient. Patient characteristics were provided at baseline by CKD stage (3a, 3b, and 4–5), and by amount of change in serum bicarbonate during the follow-up period (decrease of >2 mEq/l/yr; decrease of ≥0.5 to ≤2 mEq/l/yr; stable, defined as ±0.5 mEq/l/yr; increase of 0.5 to 2 mEq/l/yr, increase of 2 to 4 mEq/l/yr, and increase of 4+ mEq/l/yr).
The primary outcome of RRT40 was evaluated with unadjusted and adjusted time-dependent Cox proportional hazards models, with censoring at death or last date of EHR data. Death as a competing risk was examined with a Fine and Gray model. Primary exposure variables in these models were the change in serum bicarbonate from baseline, evaluated at each postindex outpatient serum bicarbonate test before event, death, or censoring as a time-dependent continuous variable, and baseline serum bicarbonate as a continuous variable. Other covariates were sex, key comorbidities as categorical variables, as well as age and eGFR as continuous variables. The proportional hazards assumption was tested by visual inspection of Schoenfeld residuals. Martingale residual plots were reviewed to assess the functional form of continuous variables. An eGFR-time interaction (baseline eGFR × Log [time]) was added as a model refinement in response to visual interpretation of the Schoenfeld residual plot. The percentage change in risk of RRT40 associated with each 1 mmol/l within-patient increase in serum bicarbonate over time was calculated as 1 minus the HR.
Sensitivity Analysis
To further adjust for potential residual confounding, a sensitivity analysis was conducted adding variables for concomitant medications, weight change and new or inpatient heart failure in a subset of the population with available data. Concomitant medications (sodium bicarbonate, metolazone, loop diuretics, and sacubitril-valsartan) were evaluated on a binary basis according to whether an outpatient prescription was identified during the observation period. Weight change was defined as the percentage change in weight from index to the last value for weight during the observation period. Inpatient or new onset heart failure was evaluated during the observation period as an inpatient admission with a concurrent diagnosis of heart failure or a diagnosis code for heart failure during the observation period in a patient without comorbid heart failure at index.
All statistical analyses were performed using SAS/STAT software, version 9.4 (Cary, NC, USA). P values <0.05 were considered statistically significant.
Results
Study Cohort and Characteristics
Of the 103 million records available in the Optum’s deidentified Integrated Claims-Clinical data set, 891,764 met the criteria for exclusion in the data extract based on the definition of advanced CKD and presence of a concurrent serum bicarbonate test described above. After applying study inclusion parameters, a total of 24,384 patients with CKD stages G3 to G5 and metabolic acidosis at index were included in the study cohort (Figure 1). Median follow-up was 3.7 years (maximum 11.5 years). Mean age was 64.9 years, 48% were male, mean index serum bicarbonate was 19.2 mmol/l and mean index eGFR was 36.9 ml/min per 1.73 m2 (Table 1). On average, patients had 19 serum bicarbonate tests (median 10 tests, interquartile range 3–24) during the outcome period. Fifty-seven percent of individuals had diagnosed hypertension, 35% had diabetes, and 25% had coronary artery disease.
Figure 1.
Study cohort selection diagram. CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate.
Table 1.
Demographics and clinical characteristics of the study cohort
| Characteristic | Baseline CKD stage |
|||
|---|---|---|---|---|
| Study cohort (N = 24,384) | CKD stage 3a (n = 8906) | CKD stage 3b (n = 7216) | CKD stage 4–5 (n = 8262) |
|
| Sex, n (%) | ||||
| Male | 11,758 (48) | 4431 (50) | 3428 (48) | 3899 (47) |
| Age (yr), mean ± SD | 64.9 ± 14.1 | 64.0 ± 14.2 | 66.1 ± 13.8 | 65.0 ± 14.2 |
| Comorbidities, n (%) | ||||
| Atrial fibrillation | 2878 (12) | 1168 (13) | 908 (13) | 802 (10) |
| Coronary artery disease | 6206 (25) | 2502 (28) | 1850 (26) | 1854 (22) |
| Diabetes | 8594 (35) | 3358 (38) | 2510 (35) | 2726 (33) |
| Heart failure | 4621 (19) | 1722 (19) | 1376 (19) | 1523 (18) |
| Hypertension | 13,891 (57) | 5333 (60) | 4098 (57) | 4460 (54) |
| Peripheral vascular disease | 5423 (22) | 2208 (25) | 1662 (23) | 1553 (19) |
| Stroke | 3039 (12) | 1184 (13) | 928 (13) | 927 (11) |
| Baseline labs, mean ± SD | ||||
| eGFR (ml/min per 1.73 m2) | 36.9 ± 15.2 | 52.8 ± 4.4 | 37.7 ± 4.3 | 18.9 ± 7.2 |
| Serum bicarbonate (mEq/l) | 19.2 ± 2.0 | 19.6 ± 1.8 | 19.3 ± 1.9 | 18.7 ± 2.3 |
| Albumin-creatinine ratio, urinary (mg/g) | 208 ± 599 | 140 ± 438 | 191 ± 603 | 324 ± 759 |
| Serum calcium, corrected (mg/dl) | 9.4 ± 0.6 | 9.4 ± 0.6 | 9.4 ± 0.6 | 9.3 ± 0.6 |
| Serum albumin (g/dl) | 3.6 ± 0.7 | 3.7 ± 0.7 | 3.6 ± 0.7 | 3.6 ± 0.7 |
| Hemoglobin (g/dl) | 12.1 ± 2.1 | 12.3 ± 2.1 | 12.0 ± 2.1 | 11.8 ± 2.1 |
| Serum potassium (mEq/l) | 4.3 ± 0.6 | 4.3 ± 0.6 | 4.3 ± 0.6 | 4.4 ± 0.6 |
| Serum phosphorous (mg/dl) | 3.7 ± 1.0 | 3.5 ± 0.9 | 3.6 ± 0.9 | 4.0 ± 1.1 |
CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate.
CKD stage 3a = baseline eGFR 45 to <60 ml/min per 1.73 m2; CKD stage 3b = baseline eGFR 30 to <45 ml/min per 1.73 m2; CKD stage 4–5 = baseline eGFR <30 ml/min per 1.73 m2.
When considering descriptive statistics by the annualized change in serum bicarbonate during the follow-up period, we found that patients with the highest increase in serum bicarbonate (≥4 mEq/l/yr) had the highest rate of heart failure at baseline (21% vs. 15% in those with a stable serum bicarbonate during follow-up). Second, patients with the greatest decrease in serum bicarbonate (any decrease) had higher rates of sodium bicarbonate prescription at baseline (19%–22% vs. 16% in those with stable serum bicarbonate, and 10%–11% in those with increasing serum bicarbonate), likely being treatment that was indicated for worsening metabolic acidosis. Other characteristics were similar across serum bicarbonate change classes (Supplementary Table S2).
Incidence of Adverse Renal and Fatal Outcomes
A total of 8866 patients (36%) experienced the composite end point RRT40; 6957 (29%) experienced kidney failure, and 5387 (22%) experienced a ≥40% decline in eGFR. The RRT40 composite outcome occurred at a rate of 9.1 events per 100 patient-years, the rate of kidney failure was 7.1 per 100 patient-years, and the rate of the ≥40% decline in eGFR was 5.5 per 100 patient-years (Figure 2). On average, patients progressed to the composite outcome in 1.7 years from their index date. A total of 12,515 deaths occurred during the outcome period, representing 51% of the patient cohort.
Figure 2.
Adverse kidney outcome event rates per 100 patient-years in patients with CKD stages G3−G5 and metabolic acidosis. RRT40 or a confirmed ≥40% decline in eGFR from baseline. eGFR, estimated glomerular filtration; RRT, renal replacement therapy.
Effect of Serum Bicarbonate and Change in Serum Bicarbonate on Renal Outcomes
A within-patient increase in serum bicarbonate over time was associated with a lower risk of the composite kidney outcome. The unadjusted HRs per 1 mmol/l increase in serum bicarbonate were 0.911 (95% CI: 0.905–0.917) for the time-dependent variable, and 0.917 (95% CI: 0.906–0.927) for baseline serum bicarbonate, both P < 0.001. After adjustment for baseline eGFR, comorbidities, and other covariates, the HR per 1 mmol/l increase in serum bicarbonate over time was largely unchanged as follows: 0.916 (95% CI: 0.910–0.922; P < 0.001), translating to an 8.4% reduction in the risk of RRT40 per 1 mmol/l increase in serum bicarbonate over time. With respect to baseline serum bicarbonate, after the above adjustments, the HR per 1 mmol/l increase in baseline serum bicarbonate increased (relative to the unadjusted analysis) to 0.973 (95% CI: 0.961–0.984). Other variables associated with reduced risk of RRT40 included increasing age; higher baseline eGFR; and having baseline coronary artery disease, evidence of prior stroke, or atrial fibrillation. Variables associated with higher risk of the composite outcome included male sex, diabetes, heart failure, and the eGFR-time adjustment, indicating that the reduction in risk of RRT40 associated with baseline eGFR moderated slightly over time (HR 1.003 [95% CI: 1.002–1.004]) (Figure 3).
Figure 3.
Hazard ratios associated with analysis of composite outcome: renal replacement therapy or ≥40% decline in eGFR. Hazard ratios were derived from an adjusted Cox proportional hazards model. eGFR, estimated glomerular filtration rate.
Sensitivity Analysis Examining Death as a Competing Risk
The association of time-dependent serum bicarbonate with the composite kidney outcome was only modestly changed after assessment of death as a competing risk, HR 0.918 (95% CI: 0.911–0.925, P < 0.001) with adjustment for the same set of covariates, translating to an 8.2% reduction in the risk of RRT40 per 1 mmol/l increase in serum bicarbonate over time (Figure 4).
Figure 4.
Hazard ratios associated with analysis of composite outcome: renal replacement therapy or ≥40% decline in eGFR, with death as a competing risk. Hazard ratios were derived from an adjusted Cox proportional hazards model. eGFR, estimated glomerular filtration rate.
Sensitivity Analysis Examining Weight Change, Heart Failure, and Concomitant Medications
In a sensitivity analysis of associations with the composite kidney outcome that added observation period covariates of percentage weight change, heart failure admission or new onset heart failure, and specified concomitant medication use (including alkali supplements), the HR associated with time-dependent serum bicarbonate was 0.911 (95% CI: 0.904–0.919, P < 0.001) (Supplementary Table S3).
Discussion
In this retrospective cohort study of 24,384 patients with CKD stages G3−G5 and metabolic acidosis, we found that both the change in serum bicarbonate over time and baseline serum bicarbonate were associated with the outcome of kidney failure or RRT40 independent of differences in baseline eGFR. Similar to the magnitude of effect of higher baseline serum bicarbonate values, we found that a 1 mmol/l increase in serum bicarbonate change over time was associated with an 8.2% to 8.4% reduction in the risk of RRT40. These findings complement an existing study that found a 7% increase in the risk of a composite renal outcome (initial of RRT, halving of eGFR, or a 25 ml/min per 1.73 m2 decline from baseline), and extend the existing literature by providing these results in a larger, community-based cohort.14
Previous studies examining a single serum bicarbonate measurement in time have shown conflicting results. An analysis of 3939 participants from the Chronic Renal Insufficiency Cohort study cohort found that a higher baseline serum bicarbonate measurement was associated with lower risk of kidney disease progression, defined as a 50% decline in eGFR or kidney failure (HR of 0.91 per 1 mmol/l increase among individuals with an eGFR <45 ml/min per 1.73 m2).11 These findings were confirmed in another analysis of the Optum data, which found that a 1 mmol/l higher baseline serum bicarbonate was associated with reduced risk of 40% decline in eGFR or kidney failure (HR 0.93 per 1 mmol/l increase among individuals with CKD stages G3–G5).13 However, these findings were not consistent in the entire literature. For example, one analysis of the Modification of Diet in Renal Disease cohort found no association with serum bicarbonate after adjustment for baseline eGFR on kidney failure and mortality outcomes.12 Though these studies provide a longitudinal assessment of outcomes, they have relied on a single measure of serum bicarbonate with long follow-up times, thereby limiting the capture of the full trajectory of the patient with respect to exposure to retained acid (metabolic acidosis).
There are several possible reasons why serum bicarbonate values may increase or decrease over time. As kidney function declines, patients progressively lose the ability to excrete acid, and serum bicarbonate levels decline. This has been well-described in studies that examine the prevalence of metabolic acidosis in CKD stages G3 to G5, and show a more than 2-fold increase in prevalence in CKD Stage G5 compared to stage G3.13 Conversely, in response to the lower serum bicarbonate values, physicians may ask patients to reduce the intake of high acid foods (e.g., meat and cheese) or prescribe oral sodium bicarbonate therapy.19 Alternatively, diuretics which are prescribed to manage volume status can lead to contraction alkalosis, and thereby raise the serum bicarbonate level.20 In addition, if patients are selected as having metabolic acidosis and enter a cohort based on serum bicarbonate values, regression to mean can result in an increased bicarbonate level in follow-up. Given that diuretics or regression to mean do not impact progression to CKD,21 our findings would suggest that either improvement of kidney function or a therapeutic effect of the interventions is likely responsible for the observed benefit. Several single-center open-label studies19 and 1 multicenter randomized double-blind placebo-controlled trial,22 have suggested the benefit of treating metabolic acidosis.
There are important clinical and research implications of our findings. Clinically, physicians often wait for more than 1 bicarbonate value in the acidosis range to initiate treatment, although all patients with metabolic acidosis are at high risk for progression. These findings suggest that a patient with a declining serum bicarbonate may be at even greater risk and should be treated promptly. Second, despite being recommended as a routine test, bicarbonate testing is often not done for patients with CKD, particularly in countries where total carbon dioxide is not a routinely measured analyte in a chemistry panel and therefore must be specially ordered. These findings suggest that serum bicarbonate should be routinely monitored, even in patients with eGFRs as high as 60 ml/min per1.73 m2 at follow-up visits, because the trajectory may be informative. From a research perspective, these findings further support the conduct of large rigorous double-blinded placebo-controlled randomized controlled trials of interventions that raise serum bicarbonate and treat metabolic acidosis in patients with CKD. There are currently studies of these therapies underway, and they should report on CKD-related outcomes in the coming years.23
This study had several strengths. First, we performed the analysis in a large, community-based US CKD population, with a follow-up period that exceeds 10 years, including patients from all types of insurance statuses, and included individuals from all US states and Puerto Rico. Second, the findings described here are perhaps more intuitive than the association of a baseline measurement, because they capture the change over time and incorporated an average of 19 measurements (median 10) with a median follow-up of nearly 4 years.
There were several limitations to this analysis. First, we relied on the capture of laboratory values during the natural course of patient care, rather than at prespecified times, and therefore may have been biased toward patients with a greater burden of illness. Second, it is possible that there is some residual confounding that was not accounted for in our models, however our sensitivity analysis in a subset of the population with concomitant medications including sodium bicarbonate and loop diuretics and change in weight (as a surrogate for volume status) demonstrated that our baseline findings were largely unchanged. Nevertheless, we were not able to account for the use of angiotensin-converting enzyme inhibitors, angiotensin-2 receptor blockers, or thiazide diuretics. Third, we were unable to control for any dietary changes or over-the-counter sodium bicarbonate usage not recorded in the medical record that may have led to changes in serum bicarbonate over time. This limitation may not have significantly affected our findings because studies from Canada, where prescription of sodium bicarbonate is the norm (vs. over-the-counter use) and from clinical studies in which all sodium bicarbonate usage was collected,11,22 suggest that a very small percentage of patients (∼2%–10%) with metabolic acidosis are treated with sodium bicarbonate (or other alkali supplement). Furthermore, adherence to both very low protein diets and to sodium bicarbonate treatment is low.24,25 Lastly, we relied on diagnosis codes for the ascertainment of covariates and not clinically adjudicated diagnoses, and thus may have not fully controlled for the respective conditions.
In conclusion, in a real-world population of over 24,000 US patients with CKD and metabolic acidosis, our results demonstrate that a within-patient increase in serum bicarbonate over time is associated with a lower risk of progression of CKD. These findings suggest that both the absolute serum bicarbonate as well as its trajectory should be considered in the decision to treat metabolic acidosis.
Disclosure
NT, TWF, NLR, JL, SEF, and VMathur were paid consultants to Tricida, Inc. in connection with the development of this manuscript. NT, NLR, and VM report consultancy, personal fees, and equity ownership from Tricida, Inc., related to the submitted work. TWF, JL, and SEF report consultancy and personal fees from Tricida, Inc. NT, VM are members of advisory boards at Tricida. VM is listed on patents related to work for Tricida. VM reports additional consulting fees from Tricida, Equillium, Myovant, Rigel, Corvidia, Acuta, Frazier, Intarcia, PTC Bio, Escient, Galderma, and Sanifit outside the submitted work.
Acknowledgments
The authors would like to thank Dr. Dawn Parsell and Dr. Jun Shao (both employees of Tricida) for review of the manuscript. Design of figures and editorial support were provided by Dr. Jun Shao. The work was partially published in abstract form at National Kidney Foundation 2021 Spring Clinical Meeting (virtual, April 6-10, 2021). This study was funded by Tricida, Inc.
Data Sharing Statement
The data that support the findings of this study are available from Optum but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available.
Author Contributions
Research idea and study design were by NT, NLR, SEF, and VM; data acquisition was done by NLR, and SEF; data analysis/interpretation was conducted by NT, TWF, NLR, JL, SEF, and VM; and statistical analysis was conducted by TWF, NLR, JL, SEF.
Each author contributed important intellectual content during manuscript drafting or revision, accepts personal accountability for the author’s own contributions, and agrees to ensure that questions pertaining to the accuracy or integrity of any portion of the work are appropriate investigated and resolved.
Footnotes
Table S1. Sources and definitions.
Table S2. Patient characteristics by annualized change in serum bicarbonate from index to last test before RRT40 outcome, death or censoring.
Table S3. Adjusted Cox proportional hazards model on renal replacement therapy or ≥40% decline in eGFR in sensitivity analysis examining weight change, heart failure, and concomitant medications.
Supplementary Material
Table S1. Sources and definitions.
Table S2. Patient characteristics by annualized change in serum bicarbonate from index to last test before RRT40 outcome, death or censoring.
Table S3. Adjusted Cox proportional hazards model on renal replacement therapy or ≥40% decline in eGFR in sensitivity analysis examining weight change, heart failure, and concomitant medications.
References
- 1.Kopple J.D., Kalantar-Zadeh K., Mehrotra R. Risks of chronic metabolic acidosis in patients with chronic kidney disease. Kidney Int Suppl. 2005;(95):S21–S27. doi: 10.1111/j.1523-1755.2005.09503.x. [DOI] [PubMed] [Google Scholar]
- 2.Kraut J.A., Madias N.E. Consequences and therapy of the metabolic acidosis of chronic kidney disease. Pediatr Nephrol. 2011;26:19–28. doi: 10.1007/s00467-010-1564-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Chen W., Abramowitz M.K. Metabolic acidosis and the progression of chronic kidney disease. BMC Nephrol. 2014;15:55. doi: 10.1186/1471-2369-15-55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Frassetto L.A., Todd K.M., Morris R.C., Jr., Sebastian A. Estimation of net endogenous noncarbonic acid production in humans from diet potassium and protein contents. Am J Clin Nutr. 1998;68:576–583. doi: 10.1093/ajcn/68.3.576. [DOI] [PubMed] [Google Scholar]
- 5.Scialla J.J., Appel L.J., Astor B.C., et al. Estimated net endogenous acid production and serum bicarbonate in African Americans with chronic kidney disease. Clin J Am Soc Nephrol. 2011;6:1526–1532. doi: 10.2215/CJN.00150111. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Wesson D.E., Simoni J. Acid retention during kidney failure induces endothelin and aldosterone production which lead to progressive GFR decline, a situation ameliorated by alkali diet. Kidney Int. 2010;78:1128–1135. doi: 10.1038/ki.2010.348. [DOI] [PubMed] [Google Scholar]
- 7.Wesson D.E., Simoni J., Broglio K., Sheather S. Acid retention accompanies reduced GFR in humans and increases plasma levels of endothelin and aldosterone. Am J Physiol Ren Physiol. 2011;300:F830–F837. doi: 10.1152/ajprenal.00587.2010. [DOI] [PubMed] [Google Scholar]
- 8.Wesson D.E., Jo C.H., Simoni J. Angiotensin II receptors mediate increased distal nephron acidification caused by acid retention. Kidney Int. 2012;82:1184–1194. doi: 10.1038/ki.2012.267. [DOI] [PubMed] [Google Scholar]
- 9.Wesson D.E., Buysse J.M., Bushinsky D.A. Mechanisms of metabolic acidosis-induced kidney injury in chronic kidney disease. J Am Soc Nephrol. 2020;31:469–482. doi: 10.1681/ASN.2019070677. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Shah S.N., Abramowitz M., Hostetter T.H., Melamed M.L. Serum bicarbonate levels and the progression of kidney disease: a cohort study. Am J Kidney Dis. 2009;54:270–277. doi: 10.1053/j.ajkd.2009.02.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Dobre M., Yang W., Chen J., et al. Association of serum bicarbonate with risk of renal and cardiovascular outcomes in CKD: a report from the chronic renal insufficiency cohort (CRIC) study. Am J Kidney Dis. 2013;62:670–678. doi: 10.1053/j.ajkd.2013.01.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Menon V., Tighiouart H., Vaughn N.S., et al. Serum bicarbonate and long-term outcomes in CKD. Am J Kidney Dis. 2010;56:907–914. doi: 10.1053/j.ajkd.2010.03.023. [DOI] [PubMed] [Google Scholar]
- 13.Tangri N., Reaven N.L., Funk S.E., Ferguson T.W., Collister D., Mathur V. Metabolic acidosis is associated with increased risk of adverse kidney outcomes and mortality in patients with non-dialysis dependent chronic kidney disease: an observational cohort study. BMC Nephrol. 2021;22:185. doi: 10.1186/s12882-021-02385-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Dobre M., Yang W., Pan Q., et al. Persistent high serum bicarbonate and the risk of heart failure in patients with chronic kidney disease (CKD): a report from the chronic renal insufficiency cohort (CRIC) study. J Am Heart Assoc. 2015;4 doi: 10.1161/JAHA.114.001599. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Raphael K.L., Zhang Y., Ying J., Greene T. Prevalence of and risk factors for reduced serum bicarbonate in chronic kidney disease. Nephrol (Carlton) 2014;19:648–654. doi: 10.1111/nep.12315. [DOI] [PubMed] [Google Scholar]
- 16.Ikizler H.O., Zelnick L., Ruzinski J., et al. Dietary acid load is associated with serum bicarbonate but not insulin sensitivity in chronic kidney disease. J Ren Nutr. 2016;26:93–102. doi: 10.1053/j.jrn.2015.08.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Wallace P.J., Shah N.D., Dennen T., Bleicher P.A., Crown W.H. Optum Labs: building a novel node in the learning health care system. Health Aff (Millwood) 2014;33:1187–1194. doi: 10.1377/hlthaff.2014.0038. [DOI] [PubMed] [Google Scholar]
- 18.Levey A.S., Stevens L.A., Schmid C.H., et al. A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150:604–612. doi: 10.7326/0003-4819-150-9-200905050-00006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Navaneethan S.D., Shao J., Buysse J., Bushinsky D.A. Effects of treatment of metabolic acidosis in CKD: a systematic review and meta-analysis. Clin J Am Soc Nephrol. 2019;14:1011–1020. doi: 10.2215/CJN.13091118. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Brinkman JE, Sharma S. Physiology, metabolic alkalosis. In: StatPearls [Internet]. StatPearls Publishing; 2022. Accessed January 20, 2023. https://www.ncbi.nlm.nih.gov/books/NBK482291/ [PubMed]
- 21.Fitzpatrick J.K., Yang J., Ambrosy A.P., et al. Loop and thiazide diuretic use and risk of chronic kidney disease progression: a multicentre observational cohort study. BMJ Open. 2022;12 doi: 10.1136/bmjopen-2021-048755. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Wesson D.E., Mathur V., Tangri N., et al. Long-term safety and efficacy of veverimer in patients with metabolic acidosis in chronic kidney disease: a multicentre, randomised, blinded, placebo-controlled, 40-week extension. Lancet. 2019;394:396–406. doi: 10.1016/S0140-6736(19)31388-1. [DOI] [PubMed] [Google Scholar]
- 23.Evaluation of effect of TRC101 on progression of chronic kidney disease in subjects with metabolic acidosis. ClinicalTrials.gov. https://clinicaltrials.gov/ct2/show/study/NCT03710291x Accessed January 30, 2023.
- 24.Garneata L., Stancu A., Dragomir D., Stefan G., Mircescu G. Ketoanalogue-supplemented vegetarian very low-protein diet and CKD progression. J Am Soc Nephrol. 2016;27:2164–2176. doi: 10.1681/ASN.2015040369. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Whitlock RH, Ferguson TW, Komenda P, et al. Metabolic acidosis is undertreated and underdiagnosed: a retrospective cohort study. Nephrol Dial Transplant. 2022:gfac099. doi: 10.1093/ndt/gfac299. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.