CRIC Study Subanalysis: Predictors of Net Acid Excretion

Due to an imbalance of acid load and excretion, metabolic acidosis is a common complication of chronic kidney disease (CKD). It is possible that higher acid load is the mechanism that links metabolic acidosis with poor kidney outcomes, due, in part, to associations between higher diet-derived acid load and faster progression of CKD.

The gold standard measure of acid load is net acid excretion (NAE). In unexpected results, higher NAE has been associated with slower CKD progression. Differences in NAE may reflect differences in diet-derived acid load in part; kidney and tubular function, body size, or metabolic acid production unrelated to dietary intake may also be involved.

Researchers have found that the associations between higher NAE and slower CKD progression are particularly pronounced in patients with diabetes mellitus, suggesting that excess acid may be produced during the altered energy metabolism characteristic of diabetes and its precursor, metabolic syndrome. Landon Brown, MD, and colleagues recently conducted a cross-sectional study to explore predictors of NAE in patients enrolled in the CRIC (Chronic Renal Insufficiency Cohort) study. Results of the current analysis were reported in the American Journal of Kidney Diseases [2019;74(2):203-212].

Candidate predictors were examined across a set of prespecified domains, including demographics, comorbid conditions, medications, laboratory values, diet, physical activity, and body composition. Each predictor was evaluated for an association with NAE in unadjusted and minimally adjusted linear regression models.

Participants for the current analysis were randomly selected from CRIC participants with 24-hour urine samples who participated in the CRIC mineral metabolism substudy (n=1000). Following obtainment of NAE measurements, 22 participants were excluded, resulting in an analysis cohort of 978 participants.

Mean age of the cohort was 58 years, 56.5% were men, 40.6% were non-Hispanic whites, and 43.5% were non-Hispanic blacks. Mean estimated glomerular filtration rate (eGFR) was 44 mL/min/1.73 m2 and ~51% had diabetes mellitus. Mean NAE was 33.2 mEq/d. NAE was higher among those with diabetes and greater levels of eGFR, insulin resistance, potential renal acid load (PRAL), and fat-free body mass.

In unadjusted analyses, characteristics associated with higher NAE included non-Hispanic white race, male sex, younger age, larger body size, greater physical activity and dietary intake, greater eGFR, higher serum albumin level, history of diabetes mellitus, increasing insulin resistance, and use of certain metabolically active medications. Following multivariable adjustment for age, sex, race, eGFR, and body surface area, the association between higher NAE and all measures of body composition remained.

There was a significant association between higher serum bicarbonate level and lower NAE, suggesting that low acid load was resulting in a higher steady-state bicarbonate concentration. There was no association between NAE and diuretics and other medications known to associate with steady-state bicarbonate concentrations. With the exception of metformin, there was no association between NAE and antidiabetic medications such as sulfonylureas, insulin, and thiazolidinediones.

Within the body composition domain, the largest effect size was observed for fat-free mass. Within the diet domain, PRAL and total dietary protein had similar effect sizes. The analysis examined serum uric acid level as a predictor post hoc, but there was no association between uric acid and higher NAE in univariate analysis.

To determine a full set of independent predictors, candidates from each of the domains were selected based on biologic rationale and strength of association in univariable, multivariable, and domain-specific models. Age, sex, race/ethnicity (non-Hispanic white, non-Hispanic black, and other), fat-free mass, homeostatic model assessment of insulin resistance, eGFR, 24-hour urine albumin excretion, presence of diabetes with and without use of metformin, and PRAL were included in the fully adjusted model.

Higher NAE remained directly associated with non-Hispanic white race, greater fat-free body mass, greater eGFR, higher insulin resistance, and higher PRAL. Among participants with diabetes, those using metformin had higher NAE compared with those not using metformin (P=.03).

Study limitations cited by the authors included the cross-sectional design; testing of urine specimens after long-term storage of up to 10 years, possibly affecting measurement accuracy; and basing urine measurements and inferences on a classic understanding of acid-base physiology.

In conclusion, the researchers said, “Overall, results from this study suggest that NAE is not only related to diet, but also body composition and metabolic factors, including metabolically active medications that could modify CKD risk. Interestingly, many of the established and emerging therapies that improve diabetic kidney disease outcomes also alter basal energy metabolism to increase acid production in diabetes mellitus. Metformin, a mainstay of diabetic therapy, is known to improve mortality, but also carries a rare risk for lactic acidosis. Newer therapies including sodium-glucose cotransporter 2 inhibitors also improve CKD outcomes while inducing subtle or frank ketosis. Sodium bicarbonate therapy may also promote augmented endogenous acid production, in part to protect against the development of metabolic acidosis, but effects on outcomes in diabetes are not known. We propose that differences in basal energy metabolism resulting in greater diet-independent acid production could explain our prior findings of improved kidney outcomes in diabetic patients with higher NAE and could be a unifying feature of kidney protective therapies in diabetes. Further studies are needed to validate this paradigm.”

Takeaway Points

  1. Researchers conducted an analysis of data from the CRIC study to examine independent predictors of net acid excretion (NAE) across multiple domains (demographics, comorbidities, medications, laboratory values, diet, physical activity, and body composition).
  2. In adjusted models, there were associations between NAE and insulin resistance, dietary potential renal acid load, and a variety of metabolically active medications.
  3. Higher NAE was also independently associated with body size, race/ethnicity, and estimated glomerular filtration rate.