Advertisement

Serum Bicarbonate and Graft and Patient Outcomes Among Kidney Transplant Recipients: A Retrospective Cohort Study Evaluating Changes in Serum Bicarbonate Over Time

Open AccessPublished:November 15, 2022DOI:https://doi.org/10.1016/j.xkme.2022.100573

      ABSTRACT

      Rationale & Objective

      Identification of treatable risk factors for kidney allograft failure is necessary to improve graft longevity. Metabolic acidosis with either low serum bicarbonate or normal serum bicarbonate (eubicarbonatemic metabolic acidosis) is implicated in native kidney disease progression but its effects in kidney transplant recipients are unclear.

      Study Design

      Retrospective cohort study.

      Setting & Participants

      An Integrated Claims-Clinical dataset of US patients with chronic kidney disease (estimated glomerular filtration rates <60 mL/min/1.73 m2) and serum bicarbonate data was used to generate a cohort of kidney transplant recipients with data from ≥1 year prior to and following transplantation.

      Primary Predictor

      The primary independent variable was a change in serum bicarbonate from baseline.

      Outcomes

      The primary outcomes were graft failure and all-cause mortality. The secondary outcomes were major adverse cardiac events and all-cause hospitalization.

      Analytical Approach

      We used adjusted time-dependent Cox proportional hazards models to assess the risk of graft failure, all-cause mortality, major adverse cardiac events, and time to first hospitalization.

      Results

      In this US community-based cohort of 1,915 kidney transplant recipients with a median follow-up of ∼2.5 years, each 1-mEq/L increase in serum bicarbonate was associated with significantly lower hazard of graft loss, death, major adverse cardiac events, and hospitalization by 10%, 8%, 4%, and 8%, respectively.

      Limitations

      Possible residual confounding.

      Conclusions

      In a US community-based population of kidney transplant recipients, even small incremental increases in serum bicarbonate (1 mEq/L) were significantly associated with reduced hazard of graft loss, all-cause mortality, cardiovascular events, and hospitalization. Interventional trials evaluating the potential benefits of treating metabolic acidosis in kidney transplant recipients are warranted.

      KEY WORDS

      INTRODUCTION

      In the United States, graft survival rates at 1-year and 5-years post-transplant decline from 93.4% to 72.4% and from 97.2% to 84.6% in deceased and donor transplants, respectively.
      • Wang J.H.
      • Skeans M.A.
      • Israni A.K.
      Current Status of Kidney Transplant Outcomes: Dying to Survive.
      At 10-years post-transplant, approximately 50% of deceased donor transplants and 30% of living donor allografts are no longer functional either due to graft loss or to death with a functioning graft.
      • Langewisch E.
      • Mannon R.B.
      Chronic Allograft Injury.
      The pathogenesis of graft failure is multifactorial and includes both immunologic and non-immunologic factors and is often preceded by progressive loss of graft function.
      • Langewisch E.
      • Mannon R.B.
      Chronic Allograft Injury.
      • Nankivell B.J.
      • Borrows R.J.
      • Fung C.L.
      • O'Connell P.J.
      • Allen R.D.
      • Chapman J.R.
      The natural history of chronic allograft nephropathy.
      • Van Loon E.
      • Bernards J.
      • Van Craenenbroeck A.H.
      • Naesens M.
      The Causes of Kidney Allograft Failure: More Than Alloimmunity. A Viewpoint Article.
      Indeed, the prevalence of interstitial fibrosis and tubular atrophy is as high as 60-70% on protocol biopsies after the first year.
      • Nankivell B.J.
      • Borrows R.J.
      • Fung C.L.
      • O'Connell P.J.
      • Allen R.D.
      • Chapman J.R.
      The natural history of chronic allograft nephropathy.
      Identification of treatable risk factors for progressive loss of graft function is necessary to improve graft longevity, reduce the need for re-transplantation, and to optimize patient outcomes.
      Metabolic acidosis is a process that results in net acid retention. It is not defined by any particular level of serum bicarbonate. Renal ammoniagenesis, and consequently, the ability to excrete acid decline progressively with loss of kidney function.
      • Wesson D.E.
      • Buysse J.M.
      • Bushinsky D.A.
      Mechanisms of Metabolic Acidosis-Induced Kidney Injury in Chronic Kidney Disease.
      This impairment results in demonstrable acid retention even in patients with earlier stages of chronic kidney disease (CKD) (i.e., stage 2, estimated glomerular filtration rates [eGFR] 60 - 89 mL/min/1.73 m2) whose serum bicarbonate levels are normal, a phenomenon known as eubicarbonatemic metabolic acidosis.
      • Alpern R.J.
      • Sakhaee K.
      The clinical spectrum of chronic metabolic acidosis: homeostatic mechanisms produce significant morbidity.
      ,
      • Goraya N.
      • Simoni J.
      • Sager L.N.
      • Pruszynski J.
      • Wesson D.E.
      Acid retention in chronic kidney disease is inversely related to GFR.
      Acid retention results in adaptations such as release of bone alkali, renal tubular hypertrophy, and muscle catabolism that are ultimately maladaptive, contributing to bone demineralization, renal fibrosis, and loss of muscle mass.
      • Wesson D.E.
      • Buysse J.M.
      • Bushinsky D.A.
      Mechanisms of Metabolic Acidosis-Induced Kidney Injury in Chronic Kidney Disease.
      Thus, a decrease in serum bicarbonate below the normal range is a late finding in the course of chronic metabolic acidosis in patients with chronic kidney disease.
      • Melamed M.L.
      • Raphael K.L.
      Metabolic Acidosis in CKD: A Review of Recent Findings.
      Metabolic acidosis with low serum bicarbonate (<22 mEq/L) is an independent risk factor for CKD progression
      • 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.
      and may also be a salient factor in the loss of graft function in transplanted kidneys. Bicarbonate levels within the lower end of the normal range may also be associated with higher risk. A previous study in a Korean cohort of kidney transplant recipients (KTRs) showed that metabolic acidosis 3 months after kidney transplantation was associated with an increased risk of graft loss, with risk of graft loss that reached a nadir at serum bicarbonate levels between 26-28 mEq/L.
      • Park S.
      • Kang E.
      • Park S.
      • et al.
      Metabolic Acidosis and Long-Term Clinical Outcomes in Kidney Transplant Recipients.
      However, such an analysis has not been reported from a large US kidney transplant recipient population. Here we report the association of changes in serum bicarbonate on graft loss, mortality, cardiovascular outcomes, and hospitalizations from a diverse community-based cohort of over 1,900 kidney transplant recipients from the United States.

      METHODS

      Study Design and Data Sources

      In this retrospective cohort study, we analyzed data from Optum’s de-identified Integrated Claims-Clinical dataset of US patients (2007-2019) with advanced CKD and serum bicarbonate data (≥ 3 eGFR results <60 mL/min/1.73 m2 and ≥ 3 serum bicarbonate results with at least 1 result between 12 and 29 mEq/L) (Figure 1). Data collection for the data extract included records from January 2007 through June 2019. We selected patients for this study from this data extract. Study cohort patients were required to have a kidney transplant procedure (by procedure code or Diagnosis Related Group [DRG]) during an inpatient admission, preceded by ≥ 1 year of data that included at least one serum bicarbonate value between 12 and 40 mEq/L. Inclusion further required patient survival and electronic health records (EHR) data for ≥ 1-year post-transplant, and ≥ 1 value each for eGFR and serum bicarbonate in the period 6 to 12 months after transplantation. Patients were excluded if there was evidence that another major organ other than the pancreas was transplanted before or concurrent with the kidney transplantation.
      Figure thumbnail gr1
      Figure 1Study cohort selection diagram. Activity in the database was defined by the presence of data point(s) that indicated a likely in-person interaction (e.g., measurement of vital signs, collection of laboratory specimens, emergency department discharger) with the patient (see Table S1 for full definition). Abbreviations: CKD, chronic kidney disease; CPT, current procedural terminology; DRG, diagnosis-related group; EHR, electronic health record; eGFR, estimated glomerular filtration rate; ICD, international classification of diseases
      The Optum dataset is a longitudinal clinical repository which contains over 103 million people with all types of health insurance plans, including those who are uninsured, from all 50 states in the United States and Puerto Rico. Extracted data elements utilized for this analysis were derived from inpatient and outpatient EHRs and administrative systems, including laboratory results, coded diagnoses and procedures, and provider notes extracted by natural language processing. Informed consent and institutional review board oversight were not required because the dataset included only de-identified information. Specific definitions, sources, data period activity, and data cleaning methods are summarized in Table S1.

      Variables

      Baseline characteristics and covariates were consistent with prior research
      • Djamali A.
      • Singh T.
      • Melamed M.L.
      • et al.
      Metabolic Acidosis 1 Year Following Kidney Transplantation and Subsequent Cardiovascular Events and Mortality: An Observational Cohort Study.
      and were defined during four time points and intervals: 1) at transplant; 2) during the first full year post-transplant; 3) during the baseline period starting 6 months post-transplant; and 4) during the outcome period starting 12 months post-transplant (Figure 2). Patient demographics, comorbidities, prior kidney transplant status, donor type, and induction immunosuppression were assessed as of the transplant date. Post-transplant infections, new-onset diabetes, and use of angiotensin-converting enzyme (ACE) inhibitors, angiotensin receptor blockers (ARBs), and diuretics were evaluated for a full year post-transplant, starting 6 months post-transplant for the pharmaceuticals. Other baseline variables, mean and maximum eGFR, blood pressure, and maintenance immunosuppression were defined during the baseline period 6 – 12 months post-transplant. Patients were classified by CKD stage using the mean baseline eGFR. We chose the period 6 – 12 months post-transplant to define the baseline serum bicarbonate and eGFR because graft function and acid-base balance can be unstable in the earlier post-transplantation period. Specific definitions are summarized in Table S1.
      Figure thumbnail gr2
      Figure 2Time periods for assessment of baseline variables, covariates, and outcomes. The primary predictors were baseline serum bicarbonate (assessed 7 – 12 months post-transplant) and time-dependent change in serum bicarbonate during the outcome period. Outcome variables were graft failure, all-cause mortality, MACE+ (a composite endpoint that included stroke, myocardial infarction, new onset heart failure, heart failure hospitalization in a patient with a pre-transplant history of heart failure, and cardiovascular death), and all-cause hospitalization. * Evaluated for 12 months, beginning 6 months post-transplant. Abbreviations: ACE, angiotensin-converting enzyme; ARB, angiotensin II receptor blocker; eGFR, estimated glomerular filtration rate; MACE+, major adverse cardiovascular event.
      The primary predictor variable was change in serum bicarbonate, which was calculated as the difference from baseline serum bicarbonate value to every serum bicarbonate test during the outcome period until the outcome event, death, or censoring at the end of EHR data. The number of days from index to each subsequent serum bicarbonate test was also retained for use in time-dependent modeling. Multiple test records on a single date were averaged. The secondary predictor variable was mean serum bicarbonate during the baseline period starting 6 months post-transplant.

      Outcomes

      The primary outcomes were graft failure and all-cause mortality. The secondary outcomes were major adverse cardiac events (MACE+) and all-cause hospitalization.
      Outcomes were evaluated during the outcome period starting at 1-year post-transplant (i.e., the index date) to the end of available data (Figure 2). Graft failure was defined using diagnosis and procedure codes as earliest occurrence of maintenance dialysis, re-transplantation, kidney transplant failure diagnosis or death. Death was identified by linkage to Social Security records prior to data de-identification. MACE+ was identified at the first occurrence during the outcome period of myocardial infarction (MI), stroke, new onset heart failure, a heart failure inpatient admission in a patient with a pre-transplant history of heart failure, or cardiovascular death (death within the same or next calendar month as discharge from a hospital inpatient stay with a diagnosis of heart failure, MI or stroke) (Table S1). Hospitalizations were defined as any inpatient hospitalization during the outcome period.
      Separate outcome-specific cohorts were established for the graft failure and MACE+ outcomes, in each case excluding patients with evidence of the outcome during the 1 year after transplantation.

      Statistical Analysis

      Patient characteristics were described using percentage or mean and standard deviation (SD), and were reported for each outcome-specific cohort in total and in groups by both baseline serum bicarbonate and change in serum bicarbonate from baseline to the last observation before event, death or censoring at end of data. The number and percentage of patients experiencing graft failure, death or MACE+ were reported for the post-index period. Incidence and frequency of hospitalization were reported for > 1 year and 3 years following the index date.
      Since serum bicarbonate is a repeated measures variable where the value for each patient may change, we employed change in serum bicarbonate from baseline in adjusted time-dependent Cox proportional hazards models to assess the hazard of graft failure, all-cause mortality, MACE+, and time to first hospitalization. Data for each patient were presented as multiple observations, and the time interval (number of days post-index) and change in serum bicarbonate value from baseline were evaluated for every post-index test until the outcome event, death, or censoring at the end of EHR data. The time-dependent indicator in our model assessed at each time interval whether a patient had the outcome at that point in time. The hazard at time t was evaluated on the of change in serum bicarbonate at that time, along with the other covariates. A time-dependent Fine and Gray model was used to evaluate a modified graft failure outcome excluding the mortality component, with death as a competing risk. Because a high serum bicarbonate, potentially identifying metabolic alkalosis, has been found to be associated with increased risk of heart failure,
      • 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.
      we performed sensitivity analyses for all 4 outcomes that excluded patients with serum bicarbonate ≥30 mEq/L at baseline. To assess the possible interaction between baseline serum bicarbonate and change in serum bicarbonate over time, as a further sensitivity analysis we constructed a categorical variable characterizing baseline serum bicarbonate as low (serum bicarbonate 12 to < 22 mEq/L), normal (serum bicarbonate ≥ 22 to < 30 mEq/L), or high (serum bicarbonate ≥ 30 mEq/L) and included an interaction term between the baseline category and the time-dependent change variable for all 4 outcomes. All adjusted statistical models were adjusted for every characteristic listed in Table 1 except for CKD stage. Statistical analyses were performed using SAS/STAT® software, version 9.4 (Cary, NC, USA). P values <0.05 were considered statistically significant.
      Table 1Demographics and clinical characteristics of the study cohort
      Overall Study Cohort (N = 1,915)*Graft Failure Study Cohort (N = 1,722)MACE+ Study Cohort (N = 1,410)
      Characteristics assessed at transplant
      Age (years), mean (SD)51.2 ± 16.751.1 (16.7)49.2 (17.0
      Female, n (%)758 (40)688 (40)586 (42)
      Body mass index (kg/m2), mean (SD)28.1 (6.1)28.0 (6.1)27.8 (6.2)
      Race, n (%)
      Asian128 (7)109 (6)99 (7)
      Black477 (25)419 (24)330 (23)
      Hispanic265 (14)236 (14)204 (14)
      Non-Hispanic White954 (50)872 (51)704 (50)
      Other/Unknown91 (5)86 (5)73 (5)
      Donor type, n (%)
      Living donor458 (24)438 (25)350 (25)
      Cadaver donor1003 (52)909 (53)726 (51)
      Unknown454 (24)375 (22)334 (24)
      Prior Transplant552 (29)462 (27)400 (28)
      Comorbidities, n (%)
      Diabetes951 (50)845 (49)646 (46)
      Hypertension1777 (93)1595 (93)1302 (92)
      Glomerulonephritis860 (45)768 (45)648 (46)
      Cystic kidney disease381 (20)348 (20)310 (22)
      Induction immunosuppression, n (%)
      Basiliximab350 (18)327 (19)257 (18)
      Anti-thymocyte globulin (rabbit)971 (51)880 (51)719 (51)
      Alemtuzumab157 (8)146 (8)136 (10)
      Other/Unknown437 (23)369 (21)298 (21)
      Smoking status, n (%)
      Current158 (8)140 (8)123 (9)
      Former590 (31)534 (31)400 (28)
      Never914 (48)820 (48)675 (48)
      Unknown253 (13)228 (13)212 (15)
      Characteristics assessed during baseline period (6-12 months post-transplant)
      Baseline serum bicarbonate (mEq/L), mean (SD)24.5 (2.8)24.6 (2.8)24.5 (2.8)
      Baseline eGFR (mL/min/1.73 m2), mean (SD)60.6 (32.6)63.4 (31.5)63.3 (33.8)
      Maximum eGFR (mL/min/1.73 m2), mean (SD)70.4 (36.3)73.2 (35.1)73.1 (37.7)
      CKD stage, n (%)
      Stage 1220 (11)209 (12)182 (13)
      Stage 2618 (32)596 (35)479 (34)
      Stage 3a525 (27)500 (29)389 (28)
      Stage 3b318 (17)281 (16)219 (16)
      Stage 4135 (7)100 (6)88 (6)
      Stage 599 (5)36 (2)53 (4)
      Systolic blood pressure, n (%)
      <120 mm Hg364 (19)337 (20)289 (20)
      120 – 132 mm Hg535 (28)474 (28)397 (28)
      133 – 144 mm Hg464 (24)409 (24)314 (22)
      >144 mm Hg327 (17)288 (17)228 (16)
      Blood pressure data missing225 (12)214 (12)182 (13)
      Diastolic blood pressure, n (%)
      <70 mm Hg562 (29)490 (28)377 (27)
      70-78 mm Hg580 (30)522 (30)416 (30)
      79-85 mm Hg358 (19)331 (19)279 (20)
      >85 mm Hg190 (10)165 (10)156 (11)
      Maintenance immunosuppression, n (%)**
      Tacrolimus1603 (84)1479 (86)1196 (85)
      Mycophenolate sodium316 (17)300 (17)216 (15)
      Mycophenolate mofetil1253 (65)1150 (67)942 (67)
      Cyclosporine69 (4)59 (3)36 (3)
      Characteristics assessed for one year post-transplant
      ACE inhibitors or ARBs prescription, n (%)**525 (27)464 (27)366 (26)
      Diuretics prescription, n (%)**764 (40)671 (39)515 (37)
      History of new onset diabetes, n (%)228 (12)206 (12)160 (11)
      History of post-transplant infection, n (%)1330 (69)1171 (68)938 (67)
      * Mortality and hospitalization were assessed in overall study cohort.
      ** Evaluated for 12 months, beginning 6 months post-transplant
      Abbreviations: ACE, angiotensin-converting enzyme; ARB, angiotensin II receptor blocker; CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate (calculated by the Chronic Kidney Disease Epidemiology Collaboration [CKD-EPI] equation); MACE+, major adverse cardiovascular event; n, number; SD, standard deviation.

      RESULTS

      Baseline Characteristics

      A total of 1,915 KTRs qualified for the full study cohort. Outcome-specific cohorts varied in size: 1,410 patients in the MACE+ outcome group which excluded those with a MACE+ event during the first-year post-transplant, 1,722 patients in the graft failure group after those with graft failure during the first-year post-transplant were excluded, and 1,915 in the all-cause mortality and hospitalization outcomes group which had no exclusions for pre-index events (Table 1).
      Demographic and clinical characteristics were similar across all outcome-specific cohorts (Table 1) as well as within cohort among patients whose bicarbonate increased, was unchanged, or increased (Tables S2, S3, and S4). The mean age ranged from 49 to 51 years, approximately 40% were female, and approximately half had diabetes. In each outcome-specific cohort, approximately 45% of the cohort was a race or ethnicity other than non-Hispanic White: approximately 24% of patients were Black, 14% were Hispanic, and approximately 7% were Asian. The ratio of deceased donors to living donors was 2:1 and approximately 28% of patients had received a prior kidney transplantation.
      The mean baseline eGFR ranged from 60.6 to 63.4 mL/min/1.73 m2 across the outcome cohorts; most patients were in CKD stage 2 (35%) or stage 3 (45%) and few were in stage 5 (2%). The mean baseline serum bicarbonate was approximately 24.5 mEq/L across the outcome cohorts and 17% of patients had a serum bicarbonate <22 mEq/L.
      Among patients with a serum bicarbonate <22 mEq/L, 14%, 16%, 25%, 23%, 19% and 2% were in CKD stages 1, 2, 3a, 3b, 4 and 5, respectively. Baseline characteristics of all study cohorts are summarized by baseline serum bicarbonate ranges in Tables S5-S7. The distribution of baseline serum bicarbonate in the study population is summarized in Figure S1 and Table S8.

      Graft Failure

      A total of 674 (39%) KTRs experienced graft failure during the outcome period with a median and maximum follow-up of 2.6 years and 10 years, respectively. Adjusted Cox proportional hazards models showed that each 1 mEq/L increase in serum bicarbonate over time and 1 mEq/L higher baseline serum bicarbonate were both associated with reduced hazard of graft failure after adjusting for multiple covariates (hazard ratio [HR] 0.90, 95% CI 0.87-0.92; P < 0.001 and HR 0.95, 95% CI 0.92-0.98; P < 0.001, respectively) (Figure 3 and Table 2). Of the other covariates included in the Cox proportional hazards models, living donor (vs. deceased donor) and higher baseline eGFR were strongly associated with a reduced hazard of graft failure (both P < 0.001), while prior transplantation and Black race (vs. non-Hispanic White) were strongly associated with an increased hazard of graft failure (P = 0.005 and P = 0.004, respectively) (Table S4). Since our definition of graft failure included death with a functioning transplant, we also analyzed death as a competing risk; the observed relationships between serum bicarbonate (both change in serum bicarbonate over time and baseline serum bicarbonate) and graft failure remained significant in this analysis (Table S9). We found similar relationships in another sensitivity analysis that excluded patients with baseline serum bicarbonate ≥ 30 mEq/L (Table S10).
      Figure thumbnail gr3
      Figure 3Effects of rising serum bicarbonate and higher baseline serum bicarbonate on graft failure, mortality, MACE+, and hospitalization. A. Change in serum bicarbonate. Each 1 mEq/L increase in serum bicarbonate significantly reduced the risk of graft failure (HR: 0.90, 95% CI: 0.87, 0.92, P < 0.001, death (HR: 0.92, 95% CI: 0.88, 0.96, P < 0.001, MACE+ (HR: 0.96, 95% CI: 0.92, 0.99, P = 0.02), and first hospitalization (HR: 0.92, 95% CI: 0.90, 0.95, P < 0.001). B. Baseline serum bicarbonate. Each 1 mEq/L higher baseline serum bicarbonate significantly reduced the risk of graft failure (HR: 0.95, 95% CI: 0.92, 0.98, P < 0.001 and death (HR: 0.94, 95% CI: 0.90, 0.99, P = 0.03) but not MACE+ or first hospitalization. Abbreviations: CI, confidence interval; HR, hazard ratio; MACE+, major adverse cardiovascular event.
      Table 2Adjusted Cox proportional hazards models for graft failure, mortality, MACE+, and hospitalization
      OutcomeTime-dependent change from baseline in post-transplant serum bicarbonate, per 1-mEq/L increaseBaseline serum bicarbonate, per 1 mEq/L increase
      HR (95% CI)p-ValueHR (95% CI)p-Value
      Graft Failure0.90 (0.87, 0.92)<0.0010.95 (0.92, 0.98)<0.001
      Mortality0.92 (0.88, 0.96)<0.0010.94 (0.90, 0.99)0.03
      MACE+0.96 (0.92, 0.99)0.020.97 (0.93, 1.02)0.19
      Hospitalization0.92 (0.90, 0.95)<0.0010.98 (0.96, 1.00)0.08
      Abbreviations: CI, confidence interval; HR, hazard ratio; MACE+, major adverse cardiovascular event.
      Note: Data from other variables included in the model are shown in Table S8.

      All-Cause Mortality

      At total of 246 (13%) KTRs died during the outcome period with a median and maximum follow-up of 2.4 years and 10 years, respectively. Adjusted Cox proportional hazards models showed that each 1 mEq/L increase in serum bicarbonate over time and each 1 mEq/L higher baseline serum bicarbonate were associated with reduced hazard of death after adjusting for multiple covariates (HR 0.92, 95% CI 0.88-0.96; P < 0.001 and HR 0.94, 95% CI 0.90-0.99; P = 0.03, respectively) (Figure 3 and Table 2). Of the covariates included in the Cox proportional hazards models, higher baseline eGFR (but not higher maximum eGFR) was associated with reduced hazard of death (P < 0.001), while the only other variable strongly associated with a higher hazard of death was older age (P < 0.001) (Table S4). We found similar relationships in a sensitivity analysis that excluded patients with baseline serum bicarbonate ≥ 30 mEq/L (Table S11).

      Major Adverse Cardiac Events

      A total of 349 (25%) KTRs experienced a MACE+ event during the outcome period with a median and maximum follow-up of 2.5 years and 10 years, respectively. Adjusted Cox proportional hazards models showed that rising serum bicarbonate over time was associated with a reduced hazard of MACE+ after adjusting for multiple covariates (HR 0.96, 95% CI 0.92-0.99, P = 0.02). Baseline serum bicarbonate, however, was not significantly associated with MACE+ (Figure 3 and Table 2). We found similar relationships in a sensitivity analysis that excluded patients with baseline serum bicarbonate ≥ 30 mEq/L (Table S12).

      Hospitalization

      During the first full year following the index date, 787 (41%) KTRs were hospitalized at least once and the total number of hospitalizations during this period was 1,713. Between > 1 year and 3 years following the index date, 710 (37%) patients had been hospitalized at least once. Adjusted Cox proportional hazards models showed that rising serum bicarbonate over time was associated with a reduced hazard of hospitalization after adjusting for multiple covariates (HR 0.92, 95% CI 0.90-0.95; P < 0.001). Baseline serum bicarbonate, however, was not significantly associated with hospitalization (Figure 3 and Table 2). We found similar relationships in a sensitivity analysis that excluded patients with baseline serum bicarbonate ≥ 30 mEq/L (Table S13).

      Additional Sensitivity Analyses

      To further examine the effect of rising serum bicarbonate on graft failure, MACE+, hospitalization, and death among patients with normal or high baseline serum bicarbonate, we examined potential interactions between baseline serum bicarbonate (low (12 to < 22 mmol/L), normal (≥ 22 to < 30 mmol/L), and high ≥ 30 mmol/L and changes in serum bicarbonate. There were no significant interactions for graft failure, MACE+, or hospitalization for any comparisons (Table S14). For death, there was no significant interaction in the comparison between the baseline low and high serum bicarbonate groups; however, the interaction for the comparison between baseline low and normal serum bicarbonate groups for this endpoint was significant.

      DISCUSSION

      In this large US community-based diverse cohort of kidney transplant recipients with a median follow-up of ∼2.5 years, we found that rising serum bicarbonate was associated with a reduced hazard for graft loss and adverse patient outcomes. Specifically, each 1 mEq/L increase in serum bicarbonate was associated with a lower hazard of graft loss, all-cause mortality, major cardiovascular events, and hospitalization by 10%, 8%, 4%, and 8%, respectively, independent of demographics, baseline eGFR, use of ACE inhibitor and ARBs, donor type, prior transplantation, immunosuppression, blood pressure, diabetes status, and other comorbid conditions. Additionally, the association of rising serum bicarbonate with a lower hazard of graft loss was consistent after adjusting for death as a competing risk (Table S15).
      Based on prior findings that incidence of graft loss was lowest in patients whose 3-month post-transplant serum bicarbonate level was between 26 – 28 mEq/L,
      • Park S.
      • Kang E.
      • Park S.
      • et al.
      Metabolic Acidosis and Long-Term Clinical Outcomes in Kidney Transplant Recipients.
      we chose to examine the effects of changes in serum bicarbonate in a population of patients with a range of baseline serum bicarbonate levels and not just in those bicarbonate levels below the normal range.
      We established a post-transplantation (vs. pre-transplantation) baseline period for a number of variables including serum bicarbonate, eGFR, blood pressure, and relevant concomitant medications as the most relevant baseline rather than using a pre-transplant period when a number of these factors would have been affected by dialysis. We chose the change in serum bicarbonate over time as our primary predictor variable (rather than baseline serum bicarbonate) to better emulate the potential effects of therapeutic interventions to increase the serum bicarbonate and to make use of multiple serum bicarbonate measures over time (rather than just one or two baseline measurements) in our time-dependent Cox proportional hazards models. We did evaluate baseline serum bicarbonate as a secondary independent variable and found that it also was associated with graft loss and all-cause mortality. The findings from our primary outcomes (graft failure and mortality) are consistent with those from a multicenter study of kidney transplant recipients from South Korea that reported a significant association between metabolic acidosis with total CO2 <22 mEq/L 3 months post-transplant with an increased risk of graft loss and mortality.
      • Park S.
      • Kang E.
      • Park S.
      • et al.
      Metabolic Acidosis and Long-Term Clinical Outcomes in Kidney Transplant Recipients.
      Our study replicated their findings in a large US community-based population, reflecting the racial and ethnic diversity of the US kidney transplant population. Furthermore, our study extended their findings by quantifying the incremental risk-reduction associated with each 1 mEq/L increase in serum bicarbonate.
      Our findings suggest that rising serum bicarbonate even among patients with serum bicarbonate in the normal range was associated with better outcomes. The optimal serum bicarbonate that is associated with better outcomes may be in the upper end of the normal range. Kovesdy, et al, found an incremental hazard reduction between serum bicarbonate values between 22 mEq/L and 28 mEq/L for the composite endpoint of pre-dialysis mortality or end-stage kidney disease in the general CKD population.
      • Kovesdy C.P.
      • Anderson J.E.
      • Kalantar-Zadeh K.
      Association of serum bicarbonate levels with mortality in patients with non-dialysis-dependent CKD.
      The physiological basis for this may be that serum bicarbonate declines late in the course of acid retention because it is maintained for a period of time by mechanisms that are maladaptive in the long-term (e.g., bone demineralization and increased acid excretion per remaining nephron). Thus, rising serum bicarbonate in patients with normal baseline serum bicarbonate may represent the adequate replenishment of previously depleted blood and bone buffers and reduced energy demands on remaining nephrons.
      • Wesson D.E.
      • Buysse J.M.
      • Bushinsky D.A.
      Mechanisms of Metabolic Acidosis-Induced Kidney Injury in Chronic Kidney Disease.
      We found that rising serum bicarbonate, but not baseline serum bicarbonate, was significantly associated with future MACE+ events (de novo stroke, myocardial infarction, heart failure; heart failure hospitalization in patients with a history of heart failure; and cardiovascular death). Djamali and colleagues, using a somewhat different set of cardiovascular outcomes that included arrhythmias and procedures for ischemic events in a cohort of KTRs from the University of Wisconsin, found that metabolic acidosis (total CO2 <20 mEq/L) 1 year post-transplant was a significant predictor of increased risk of cardiovascular events.
      • Djamali A.
      • Singh T.
      • Melamed M.L.
      • et al.
      Metabolic Acidosis 1 Year Following Kidney Transplantation and Subsequent Cardiovascular Events and Mortality: An Observational Cohort Study.
      Due to the differences in outcome measures and the study population, direct comparisons cannot be made with our study, but both studies found an association of low serum bicarbonate with increased cardiovascular risk in KTRs. Compared with the University of Wisconsin population, our study population had a higher proportion of non-Hispanic Whites (45% vs. 20%), a lower proportion of recipients of living donor transplants (25% vs. 41%), and somewhat higher proportion of patients that had been re-transplanted (27% vs. 20%). Our observed associations of reduced MACE+ outcomes with rising serum bicarbonate are consistent with similar observations in the community-based CKD population.
      • Collister D.
      • Ferguson T.W.
      • Funk S.E.
      • Reaven N.L.
      • Mathur V.
      • Tangri N.
      Metabolic Acidosis and Cardiovascular Disease in CKD.
      Patients with CKD retain acid due to a reduced nephron mass that is unable to produce sufficient ammonia to excrete acid from dietary sources and metabolism. Importantly, the degree of tubular injury at any given level of eGFR is an important determinant of the ability to handle an acid load. Unlike the CKD population in which there is a monotonic increase in the prevalence of metabolic acidosis from CKD stage 2 through stages 4 and 5,
      • Inker L.A.
      • Coresh J.
      • Levey A.S.
      • Tonelli M.
      • Muntner P.
      Estimated GFR, albuminuria, and complications of chronic kidney disease.
      we found that the proportion of KTRs with a serum bicarbonate <22 mEq/L was similar among those with CKD stages 3a, 3b, and 4 (25%, 23%, and 19%, respectively) and that the prevalence of serum bicarbonate <22 mEq/L among patients with CKD stages 2 and 3a post-transplant was higher than would be expected based on data from the NHANES CKD population (16% vs. 8% for CKD stage 2 and 29% vs. 9% for CKD stage 3a). These findings are consistent with the previously described high prevalence of renal tubular acidosis in KTRs and prevalent transplant-specific etiologies of tubular injury, including calcineurin inhibitor toxicity, acute rejection, and ischemia.
      • Keven K.
      • Ozturk R.
      • Sengul S.
      • et al.
      Renal tubular acidosis after kidney transplantation--incidence, risk factors and clinical implications.
      Identification of treatable risk factors that can slow graft loss and improve patient outcomes is critical to maximizing the public health and individual patient benefits that can be reaped from the limited number of organs available for transplantation. Our study shows that rising serum bicarbonate, even in small increments (e.g., 1 mEq/L) is associated with improved graft and patient survival. Cardiovascular events are the leading cause of death among KTRs.
      • Awan A.A.
      • Niu J.
      • Pan J.S.
      • et al.
      Trends in the Causes of Death among Kidney Transplant Recipients in the United States (1996-2014).
      The reduced mortality associated with rising serum bicarbonate was likely due, in part, to the reduction in MACE+ events. It would be expected that reducing the risk of graft loss and cardiovascular events should reduce hospitalization. This expectation was borne out by our analysis of hospitalizations.
      In multiple randomized, controlled trials, treatment of patients with CKD and metabolic acidosis slowed progression of chronic kidney disease and/or improved composite outcomes of death, dialysis, or eGFR decline.
      • de Brito-Ashurst I.
      • Varagunam M.
      • Raftery M.J.
      • Yaqoob M.M.
      Bicarbonate supplementation slows progression of CKD and improves nutritional status.
      • Garneata L.
      • Stancu A.
      • Dragomir D.
      • Stefan G.
      • Mircescu G.
      Ketoanalogue-Supplemented Vegetarian Very Low-Protein Diet and CKD Progression.
      • Goraya N.
      • Simoni J.
      • Jo C.H.
      • Wesson D.E.
      A comparison of treating metabolic acidosis in CKD stage 4 hypertensive kidney disease with fruits and vegetables or sodium bicarbonate.
      • 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.
      • Di Iorio B.R.
      • Bellasi A.
      • Raphael K.L.
      • et al.
      Treatment of metabolic acidosis with sodium bicarbonate delays progression of chronic kidney disease: the UBI Study.
      • Dubey A.K.
      • Sahoo J.
      • Vairappan B.
      • Haridasan S.
      • Parameswaran S.
      • Priyamvada P.S.
      Correction of metabolic acidosis improves muscle mass and renal function in chronic kidney disease stages 3 and 4: a randomized controlled trial.
      However, no studies to date have examined the effects of correcting acidosis on graft function in kidney transplant recipients.
      Our study has a number of strengths. To our knowledge, this is the largest study of community-based KTRs to evaluate the effects of metabolic acidosis in this population. Although generalizability to U.S. population of kidney transplant recipients cannot be confirmed, it is strengthened by the fact that the population was drawn from nearly 900,000 patients with eGFR <60 mL/min/1.73 m2 and concurrent serum bicarbonate measurements who were, in turn, drawn from a pool of over 103 million patients from all 50 U.S. states with all insurance types and insurance statuses. We used both diagnostic and procedure codes to identify graft failure and cardiovascular outcomes and social security records to ascertain deaths. Our primary predictor variable, time-dependent change from baseline in serum bicarbonate, utilized all available bicarbonate values after 1-year post-transplant and prior to outcomes during a median follow up of ∼2.5 years, rather than just one or two baseline measurements.
      Limitations of the study include that it was observational and thus, despite adjustment for multiple covariates, the possibility of residual confounding exists. Further, although the database we used was large and included many healthcare systems and physician practices across all US regions, it is theoretically possible for patients to move to a healthcare system that did not partner with Optum, potentially leading to missed outcomes. This dataset was derived from a well-recognized EHR data repository and not from a transplant center or centers. As such, we did not have data on donor characteristics or detailed recipient characteristics and matching, but we included data on induction and maintenance immunosuppression. We did not capture all interventions that impact serum bicarbonate longitudinally including over-the-counter sodium bicarbonate use and dietary changes. In previous studies in non-transplant patients, however, use of sodium bicarbonate among patients with metabolic acidosis was very low.
      • 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.
      As this was not an interventional study, our findings of the association of rising serum bicarbonate with better graft and patient outcomes do not necessarily reflect benefit of treatment, particularly since the literature suggests that only a small fraction of patients with CKD receive treatment for metabolic acidosis.
      • 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.
      ,
      • 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.
      Finally, although in the setting of chronically low serum bicarbonate in an outpatient with CKD is most likely to represent metabolic acidosis, rather than another acid-base disorder, we acknowledge that some patients with low serum bicarbonate may not have had metabolic acidosis.
      In conclusion, we found that in a US community-based population of kidney transplant recipients even small incremental increases in serum bicarbonate (1 mEq/L) were significantly associated with reduced hazard of graft loss, all-cause mortality, cardiovascular events, and hospitalization. The findings from our study, when taken together with the work of others,
      • Park S.
      • Kang E.
      • Park S.
      • et al.
      Metabolic Acidosis and Long-Term Clinical Outcomes in Kidney Transplant Recipients.
      ,
      • Djamali A.
      • Singh T.
      • Melamed M.L.
      • et al.
      Metabolic Acidosis 1 Year Following Kidney Transplantation and Subsequent Cardiovascular Events and Mortality: An Observational Cohort Study.
      point to the need for well-designed interventional studies evaluating the effect of treating metabolic acidosis in kidney transplant recipients.

      Acknowledgements

      The authors would like to thank Dr. Dawn Parsell and Dr. Jun Shao for review of the manuscript. Design of figures and editorial support were provided by Dr. Jun Shao.

      REFERENCES

        • Wang J.H.
        • Skeans M.A.
        • Israni A.K.
        Current Status of Kidney Transplant Outcomes: Dying to Survive.
        Adv Chronic Kidney Dis. 2016; 23: 281-286
        • Langewisch E.
        • Mannon R.B.
        Chronic Allograft Injury.
        Clin J Am Soc Nephrol. 2021; (CJN.15590920)
        • Nankivell B.J.
        • Borrows R.J.
        • Fung C.L.
        • O'Connell P.J.
        • Allen R.D.
        • Chapman J.R.
        The natural history of chronic allograft nephropathy.
        N Engl J Med. 2003; 349: 2326-2333
        • Van Loon E.
        • Bernards J.
        • Van Craenenbroeck A.H.
        • Naesens M.
        The Causes of Kidney Allograft Failure: More Than Alloimmunity. A Viewpoint Article.
        Transplantation. 2020; 104: e46-e56
        • 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
        • Alpern R.J.
        • Sakhaee K.
        The clinical spectrum of chronic metabolic acidosis: homeostatic mechanisms produce significant morbidity.
        Am J Kidney Dis. 1997; 29: 291-302
        • Goraya N.
        • Simoni J.
        • Sager L.N.
        • Pruszynski J.
        • Wesson D.E.
        Acid retention in chronic kidney disease is inversely related to GFR.
        Am J Physiol Renal Physiol. 2018; 314: F985-F991
        • Melamed M.L.
        • Raphael K.L.
        Metabolic Acidosis in CKD: A Review of Recent Findings.
        Kidney Med. 2021; 3: 267-277
        • 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
        • Park S.
        • Kang E.
        • Park S.
        • et al.
        Metabolic Acidosis and Long-Term Clinical Outcomes in Kidney Transplant Recipients.
        J Am Soc Nephrol. 2017; 28: 1886-1897
        • Djamali A.
        • Singh T.
        • Melamed M.L.
        • et al.
        Metabolic Acidosis 1 Year Following Kidney Transplantation and Subsequent Cardiovascular Events and Mortality: An Observational Cohort Study.
        Am J Kidney Dis. 2019; 73: 476-485
        • 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
        • Kovesdy C.P.
        • Anderson J.E.
        • Kalantar-Zadeh K.
        Association of serum bicarbonate levels with mortality in patients with non-dialysis-dependent CKD.
        Nephrol Dial Transplant. 2009; 24: 1232-1237
        • Collister D.
        • Ferguson T.W.
        • Funk S.E.
        • Reaven N.L.
        • Mathur V.
        • Tangri N.
        Metabolic Acidosis and Cardiovascular Disease in CKD.
        Kidney Med. 2021; 3: 753-761
        • Inker L.A.
        • Coresh J.
        • Levey A.S.
        • Tonelli M.
        • Muntner P.
        Estimated GFR, albuminuria, and complications of chronic kidney disease.
        J Am Soc Nephrol. 2011; 22: 2322-2331
        • Keven K.
        • Ozturk R.
        • Sengul S.
        • et al.
        Renal tubular acidosis after kidney transplantation--incidence, risk factors and clinical implications.
        Nephrol Dial Transplant. 2007; 22: 906-910
        • Awan A.A.
        • Niu J.
        • Pan J.S.
        • et al.
        Trends in the Causes of Death among Kidney Transplant Recipients in the United States (1996-2014).
        Am J Nephrol. 2018; 48: 472-481
        • de Brito-Ashurst I.
        • Varagunam M.
        • Raftery M.J.
        • Yaqoob M.M.
        Bicarbonate supplementation slows progression of CKD and improves nutritional status.
        J Am Soc Nephrol. 2009; 20: 2075-2084
        • 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
        • Goraya N.
        • Simoni J.
        • Jo C.H.
        • Wesson D.E.
        A comparison of treating metabolic acidosis in CKD stage 4 hypertensive kidney disease with fruits and vegetables or sodium bicarbonate.
        Clin J Am Soc Nephrol. 2013; 8: 371-381
        • 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
        • Di Iorio B.R.
        • Bellasi A.
        • Raphael K.L.
        • et al.
        Treatment of metabolic acidosis with sodium bicarbonate delays progression of chronic kidney disease: the UBI Study.
        J Nephrol. 2019; 32: 989-1001
        • Dubey A.K.
        • Sahoo J.
        • Vairappan B.
        • Haridasan S.
        • Parameswaran S.
        • Priyamvada P.S.
        Correction of metabolic acidosis improves muscle mass and renal function in chronic kidney disease stages 3 and 4: a randomized controlled trial.
        Nephrol Dial Transplant. 2020; 35: 121-129