Uremic retention solutes: the free and the bound.

M any in the nephrology community were surprised when the Hemodialysis (HEMO) Study, completed in 2002, supported the null hypothesis for each intervention, suggesting ceiling effects for both the dose of dialysis and membrane permeability (1). Neither increases in the dose (expressed as the product of dialyzer urea clearance and dialysis time factored for patient size) nor an improvement in dialyzer membrane permeability or flux (expressed as 2microglobulin clearance) improved patient survival or seven other prespecified outcome variables. In the HEMO Study, 1846 randomized patients were followed up to 7 yr, and the study was reasonably powered to detect a mortality improvement. Although outcomes were improved in subsets of the study population, the favorable effects were only suggestive, and the magnitudes were relatively small. What does this mean? Does it show that we have gone as far as we can with improving thrice-weekly dialysis or is there more to be done? The ongoing high high rates of mortality and cardiovascular disease, as well as uncontrolled data that show improvements in both of these outcomes by increasing dialysis frequency, suggest that the job is not yet done. Nephrologists perhaps optimistically refer to dialysis as “renal replacement therapy,” and have focused measurement of dialysis adequacy on low molecular weight solute clearance, with urea as the predominant indicator solute. There is no question that small solute removal is vitally important. There is also no question that conventional thrice-weekly dialysis does not replicate the global ability of native kidneys to maintain homeostasis of the interior environment or to fully alleviate the uremic syndrome. Increasingly, data suggests that maintaining even minimal residual native kidney function in dialysis patients may be a more important determinant of patient outcome than dialyzer clearance (2). Endocrine function of the kidney is important but is unlikely to account for the magnitude of these effects. What else could account for failure of the artificial kidney to keep patients healthy, in contrast to the success of the native kidney? Over the past four decades, several explanations and alternative approaches to improving dialysis outcomes have been suggested. Initial efforts focused on molecular size, as embodied in the term “middle molecule clearance” proposed by Babb et al. (3). These early pioneers envisioned that because the dialyzers in use at that time had an absolute molecular cutoff of about 10,000 Daltons and efficiently removed compounds up to only about 3000 Daltons, more toxic compounds would likely be found in the molecular weight range where dialytic removal was significant but incomplete, i.e. 500 to 5000 Daltons. They also hypothesized that molecular size could explain the success of peritoneal dialysis, because the peritoneal membrane leaks albumin and must therefore be more permeable to middle molecules. Vigorous attempts followed to improve the permeability of commercial dialyzer membranes and to improve removal of larger solutes through these same “high-flux” membranes by filtration instead of dialysis. After technical problems were resolved, these membranes gradually replaced the older, less permeable membranes. However, despite widespread application of these more permeable membranes, improvements in patient outcomes have been meager at best, and were not evident in the HEMO study. Other potentially beneficial functions provided by the native kidney include selective reabsorption of solutes from the glomerular filtrate and active secretion of solutes from the peritubular circulation into the filtrate. Tubular secretion of solutes has long puzzled physiologists. The mechanisms, especially for secretion of organic anions, are extremely robust, such that Pitts postulated over 30 yr ago that the existence of some very toxic substances in the blood must be maintained at very low concentrations (4). Others have postulated that the vigor of the secretory mechanism is necessary to remove substances bound to circulating macromolecules, principally albumin, during its passage through the peritubular circuit (5). The low free concentrations available for diffusion through the capillary endothelium would require a robust transporter to maintain a maximal gradient for diffusion throughout the length of the peritubular capillary. The success of this postulated mechanism for stripping bound ligands from albumin has been demonstrated in vitro as well as in the isolated perfused rat kidney (6). In humans and whole animals, tubular secretion of phenol red results in 80% extraction during a single pass through the kidneys despite 90% binding to serum albumin (4). In uremic patients, albumin-binding sites are highly saturated as evidenced by relatively poor binding to exogenous ligands such as radiolabeled salicylate or phenytoin (7). Extensive studies in the late 1980s demonstrated that the binding defect is due to reversibly bound ligands. The albumin binding defect has long been recognized but relatively ignored as a potential mechaPublished online ahead of print. Publication date available at www.jasn.org.