Kinetic equations intended to explain the behavior of small molecules during a hemodialysis cycle were described more than 30 years ago using urea as an example.1 They have since been extensively debated, developed, and discussed.2 A ratio, the product of urea clearance (K, in mL/min) and dialysis session length (t, in minutes) to the volume of urea distribution in the body (V, in mL), is a key parameter in all such formulations, the Kt/V. It and a related measure, the urea reduction ratio (URR), are currently accepted measures for judging dialysis dose.3
Urea is an easily measured small molecule that is eliminated by the kidney. Blood urea concentration (BUN) rises in kidney failure, so urea was considered a convenient marker for other water-soluble, potentially toxic small molecules that are removed by dialysis. Its kinetic behavior during hemodialysis was presumed similar to that of those other small molecules.
Investigations presented elsewhere in this issue of the American Journal of Kidney Diseases show that assumption is incorrect.4 Among low molecular weight guanidino compounds, there were substantial differences in plasma concentration reduction ratios (RRp) during dialysis, ranging from 49% to 82%. The largest molecule (175 Da) was associated with the highest RRp, 82%, while the smallest molecule (59 Da) was associated with a much lower RRp, 55% (Table 4 of reference 4). The plasma extraction ratios across the dialyzer, on the other hand, were comparable for all molecular species and were relatively constant during dialysis for each, so they were comparably dialyzed (Table 5 of reference 4). Those combined observations are sufficient by themselves to prove the different kinetic behaviors among these compounds. The findings are strengthened by observed differences in the change of erythrocyte-to-plasma concentration ratios among the molecules during the course of dialysis and by previous studies from these authors showing differences in the apparent volumes of distribution among the compounds using 2-pool kinetic models.5
Do these findings mean that urea should now be abandoned as a marker for dialysis dose in favor of another substance of greater presumed clinical toxicity? I do not believe so for a number of reasons that are summarized below. The way in which small-molecule dialysis is used clinically to judge dose, however, may need to change.
The National Cooperative Dialysis Study (NCDS) was a randomized, controlled trial designed to evaluate urea-directed treatment as a marker for small-molecule removal, and dialysis session length as a proxy for the removal of larger molecules.6, 7 It used a 2 × 2 factorial design to evaluate 2 target levels of t (4.5-5 hours versus 2.5-3.5 hours) and 2 target levels of time-averaged BUN concentration (TAC; 50 mg/dL versus 100 mg/dL). A urea kinetic system was used as a tool to control TAC, but Kt/V was allowed to “float” and was not controlled, monitored, or measured directly. As expected, both TAC and Kt/V differed significantly between randomized groups.
The Patient Safety Committee recommended the study be stopped after only 151 patients had been randomized because the clinical failure rate was much higher when TAC was high than when it was low. Analysis revealed that the high TAC arms had significantly higher failure than the low TAC arms (P < 0.0001).6 Risk was marginally worse if t was short (P = 0.06; Relative Risk ≅ 2.1). The effects of TAC and t were independent because there was no significant statistical interaction between them.6 Thus, it is abundantly clear that adequate small molecule–directed therapy, even using urea as the marker, is necessary to avoid preventable risk. That observation cannot be neglected even though the data are now more than a quarter century old. Similarly, one cannot ignore the possible, though weaker, effect of t (based on a “non-significant p-value” of 0.06) given the study was stopped early.8
The Hemodialysis (HEMO) Study was the other randomized controlled trial designed to address this problem.9 It evaluated 1,846 patients and also used a 2 × 2 factorial design. Dialysis dose estimated by a Kt/V (standard versus high) and membrane flux (high versus low) were the treatment arms. The Kt/V (and URR) were measured and controlled directly while BUN was allowed to float. Again, both Kt/V and TAC differed significantly between randomized groups. However, the HEMO Study was essentially a negative study; there were no significant effects of standard versus high Kt/V or low versus high flux.9
The NCDS and HEMO Study were quite different, and 1 study cannot be used to confirm or deny the conclusions of the other. Neither are their findings necessarily incompatible. Possibly, a higher dialysis dose is more important at the lower range of TAC and Kt/V utilized in the NCDS compared to the higher range utilized in the HEMO Study. Alternatively, the difference in results may reflect each study’s view of the dialysis dose. NCDS evaluated small molecule–directed therapy using urea as a convenient marker. HEMO evaluated specific values of Kt/V for urea per se. Thus, 1 study evaluated a principle for treatment (NCDS: small molecules are important or not) while the other evaluated an implementation strategy for treatment (HEMO: kinetic systems for urea). The reconciling conclusion may be that the principle is important but not necessarily the method and there are now at least 2 facts external to those studies that support the idea.
First, the studies by Eloot and colleagues reported here4 and elsewhere5 show that the kinetic behaviors during dialysis among urea and other small molecular species are quite different even though their dialyses from plasma are comparable. Thus, the specific kinetic behavior of any 1 molecule, and therefore the equations used to describe that behavior, cannot be presumed a priori to accurately reflect, or be a proxy for, any of the others. That fact was not well understood when the urea kinetic construct was originally described1 or when either of the trials discussed here6, 9 was designed.
Second, body size, whether described as body surface area (BSA), body mass index, body weight, or V, is strongly associated with survival among dialysis patients.10, 11, 12, 13 This was not understood when solute kinetics during dialysis were described or when the trials were designed.1 Higher Kt is also associated with lower risk.14 Dividing Kt by V, as a Kt/V, divides 1 measure favorably associated with survival by another. That mathematical construct leads to confused interpretation when it is used to reflect an outcome like survival15 and to inadequate dialysis for small patients.13, 16 Small or malnourished persons can have both high Kt/V (or URR) and high death risk simply because they have small body size.15 Similarly, large persons can have lower Kt/V (or URR) but better survival, both associated with larger body size. Thus, it is logically advisable to separate the components of Kt/V into at least its numerator and denominator, Kt from V. Current kinetic constructs, then, while remaining valuable for evaluating dialysis dynamics, have little relevance for evaluating an outcome like survival because current models require Kt/V as an essential mathematical requirement.
Current technology now allows direct measurement of a small-molecule K, which is very close to the urea K, during the course of each and every dialysis.17 Curvilinear relationships between a Kt treatment target so measured and body size (as BSA) have been described18 and validated.19 In principle, a target Kt for the patient’s BSA could be selected, and K and t could be prescribed accordingly.19 The Kt could then be monitored during each and every treatment without need to estimate a value for V. Thus, small molecule–directed therapy could be prescribed and monitored without the technical, mathematical (eg, number of pools and intercompartmental solute movement), and presumption (eg, V is only a simple diluent for urea and not risk-related) requirements and uncertainties of current kinetic methods. Clinical targets for Kt have not been determined by a trial similar to the NCDS or HEMO studies at this time, but the findings of empirical studies18, 19 could be confirmed or amended by such a trial. The studies by Eloot and colleagues4, 5 suggest that targets for Kt may differ according to the kinetic behavior and toxicity of retained solutes but their relative relationships to urea should be relatively constant across solutes, making unnecessary the use of multiple solutes to judge treatment.
Direct measurements of K and t are not plagued by kinetic problems. The clearance relationships are similar among the small molecules4 and remains relatively constant over the length of a dialysis treatment, unlike the movement of solutes in the body.4, 5 The relationship of an externally measured small molecule Kt to clinical outcome is determined by empirical methods18, 19 and does not depend on the movement of molecules between body compartments. Therefore, one need not abandon a simple small-molecule clearance measurement as the marker for adequate dialysis, searching endlessly and hopelessly for the “best” small molecule to evaluate with kinetic equations.
The real issues here relate to the way small-molecule dialysis is viewed in the combined context of clinical outcome and treatment prescription. The kinetic equations do not include a term for clinical outcome and the kinetic behaviors of small molecules can be very different. Kinetic equations require the division of 1 measure favorably associated with survival by another—the division of Kt by V. The direct measurement of Kt during dialysis, on the other hand, is not so constrained and the association of clearance with outcome has been demonstrated.13, 14, 16, 18, 19 Therefore, kinetic-related measures should likely be used primarily in research settings to assist better understanding the complex physiology of dialysis. Treatment adequacy should be judged in the clinical setting by simpler and more certain outcome-related measures that can be generalized across a spectrum of small molecules.
FULL TEXT
