Table 1 Observational studies of serum bicarbonate and long-term outcomes in persons without end-stage renal disease
From: Metabolic acidosis and the progression of chronic kidney disease
Study | Population | Main outcome(s) | Findings | pH or pCO2available | Strengths | Limitations |
|---|---|---|---|---|---|---|
Shah et al. 2009 [4] | 5,422 outpatients in the Bronx, NY; 9% with eGFR < 60 mL/min/1.73 m2 | Kidney disease progression, defined as 50% decrease in eGFR or eGFR < 15 mL/min/1.73 m2 | HR 1.54 (95% CI 1.13-2.09) for progression, for serum bicarbonate ≤22 mEq/L compared with 25–26 mEq/L | No | • Ethnically diverse cohort | • Single measure of serum bicarbonate |
• Data derived from clinical and administrative dataset | ||||||
Menon et al. 2010 [12] | 1,781 participants (839 randomized, 942 non-randomized) from the MDRD study | (1) ESRD (need for dialysis or transplantation); (2) all-cause mortality; (3) composite of 1 and 2 | HR 1.05 (0.87-1.28), 0.99 (0.75-1.13), 1.04 (0.87-1.24) for need for kidney failure, all-cause mortality, and composite outcome, respectively, for serum bicarbonate 11–20 compared with 26–40 mEq/L | No | • Well-characterized cohort | • Single measure of serum bicarbonate |
• Adjustment for measured GFR | ||||||
Raphael et al. 2011 [11] | 1,090 participants of the AASK trial | Composite outcome of death, ESRD (dialysis or transplantation), or GFR event (defined as a GFR reduction by 50% or by 25 ml/min/1.73 m2 from baseline) | HR 0.960 (0.924-0.998) for composite outcome, per mEq/L higher baseline serum bicarbonate | No | • Well-characterized cohort | • Single measure of serum bicarbonate |
• Adjustment for measured GFR | ||||||
• Adjustment for errors in measurement of GFR and proteinuria | ||||||
Kovesdy et al. 2009 [10] | 1,240 adults at a Veterans Affairs Medical Center; 87% with CKD stages 3 and 4 | (1) All-cause mortality; (2) composite of predialysis mortality and initiation of dialysis | U-shaped association; HR for mortality 1.43 (1.10-1.87) for serum bicarbonate <22 compared with 26–29 mEq/L; similar results for composite outcome | No | • Adjustment for time-varying serum bicarbonate levels | • Data derived from clinical and administrative dataset |
Navaneethan et al. 2011 [13] | 41,749 outpatients with eGFR < 60 mL/min/1.73 m2 in Cleveland, OH | All-cause mortality | U-shaped association; HR 1.23 (1.16-1.31) for bicarbonate <23 compared with 23–32 mEq/L; HR 1.59 (1.49-1.69) for reaching bicarbonate <23 mEq/L | No | • Examination of temporal change in serum bicarbonate | • Data derived from clinical and administrative dataset |
• Large sample size | ||||||
Dobre et al. 2013 [14] | 3,939 participants from the CRIC study | (1) Renal outcome, defined as 50% decrease in eGFR or ESRD (dialysis or transplantation); (2) atherosclerotic events; (3) CHF events; (4) all-cause mortality | Per mEq/L higher serum bicarbonate, HR 0.97 (0.94-0.99) for renal outcome; 0.99 (0.95-1.03) for atherosclerotic event; 1.14 (1.03-1.26) for CHF for serum bicarbonate ≥24 mEq/L; 0.98 (0.95-1.02) for mortality | No | • Well-characterized cohort | • Single measure of serum bicarbonate |
Kanda et al. 2013 [15] | 113 Japanese patients ≥60 years old with eGFR < 60 mL/min/1.73 m2 | Kidney disease progression, defined as 25% decrease in eGFR or initiation of dialysis | HR 0.791 (0.684-0.914) for progression, per mEq/L higher serum bicarbonate | No | • Focus on elderly cohort | • Single measure of serum bicarbonate |
• Small sample size | ||||||
Raphael et al. 2013 [16] | 15,836 participants of NHANES III | All-cause mortality | HR 1.75 (1.12-2.74), 1.56 (0.78-3.09), and 2.56 (1.49-4.38) for total population, non-CKD, and CKD subgroups, respectively, for serum bicarbonate <22 mEq/L compared with 26–30 mEq/L | No | • Nationally representative cohort | • Single measure of serum bicarbonate |
• Compared CKD and non-CKD subgroups |