Association between serum bicarbonate and death in hemodialysis patients: is it better to be acidotic or alkalotic?

The optimal acid-base status for survival in maintenance hemodialysis (MHD) patients remains controversial. According to recent reports, acidosis is associated with improved survival in MHD patients. It was hypothesized that this inverse association is due to a confounding effect of the malnutrition-inflammation complex syndrome (MICS). Associations between baseline (first 3 mo averaged) predialysis serum bicarbonate (HCO3(-)) and 2-yr mortality were examined in 56,385 MHD patients who were treated in virtually all DaVita dialysis clinics across the United States. The range of HCO3(-) was divided into 12 categories (< 17, > or = 27, and 10 groups in between). Three sets of Cox regression models were evaluated to estimate hazard ratios of all-cause and cardiovascular death in both incident and prevalent patients: (1) Unadjusted, (2) multivariate case mix adjusted (which also included dialysate HCO3(-) and Kt/V), and (3) adjusted for case mix and nine markers of MICS (body mass index; erythropoietin dose; protein intake; serum albumin; creatinine; phosphorus; calcium; ferritin and total iron binding capacity; and blood hemoglobin, WBC, and lymphocytes). There were significant inverse associations between serum HCO3(-) and serum phosphorus and estimated protein intake. The lowest unadjusted mortality was associated with predialysis HCO3(-) in the 17- to 23-mEq/L range, whereas values > or = 23 mEq/L were associated with progressively higher all-cause and cardiovascular death rates. This association, however, reversed after case-mix and MICS multivariate adjustment, so that HCO3(-) values >22 mEq/L had lower death risk. Although previous epidemiologic studies indicated an association between high serum HCO3(-) and increased mortality in MHD patients, this effect seems to be due substantially to the effect of MICS on survival.