Is an Increased Serum Bicarbonate Concentration during Hemodialysis Associated with an Increased Risk of Death?

Joline L. T. Chen,K. Kalantar-Zadeh
DOI: https://doi.org/10.1111/sdi.12231
2014-05-01
Seminars in Dialysis
Abstract:Metabolic acidosis is a common complication in chronic kidney disease (CKD) and end stage renal disease (ESRD) and is known to be associated with deleterious outcomes under these conditions.1, 2 Consistent with this biologically plausible notion, several clinical studies have suggested that correction of acidosis and/or higher bicarbonate levels are associated with greater survival in maintenance hemodialysis (MHD) patients.3–5 Hemodialysis therapy can effectively correct metabolic acidosis. During each hemodialysis session, patients are exposed to both the prescribed bicarbonate bath, usually between 30–40 mEq/L, and the dry acid concentrate, which can yield up to 8 mEq/L of bicarbonate.6, 7 Thus, a total bicarbonate load during a single dialysis session can be quite variable. A number of recent epidemiologic studies, however, have suggested that a higher bicarbonate bath concentration in dialysis and/or a higher level of achieved serum bicarbonate may paradoxically and counterintuitively be associated with an increased risk for death.8–10 These recent findings have led to an FDA advisory in 2012 and ongoing controversies and litigations.6 Hence, the safety of prescribed bicarbonate bath in MHD patients and the ideal target serum bicarbonate concentration are in the public spotlight. We believe that these paradoxical associations between higher serum bicarbonate concentration and poor outcomes are likely false and stem from the confounding role of better nutritional status and higher dietary protein intake, which result in acidemia but improved survival (see below). On the other hand, more profound metabolic acidosis, as a consequence of more severe protein-energy wasting and hypercatabolism, likely worsens survival chance and also prompts clinicians to prescribe higher bicarbonate concentration; hence, a misleading association may arise from the notorious “confounding by indication” (see below).
What problem does this paper attempt to address?