Patent Application
20100217178
The present disclosure relates to medical fluid delivery and more specifically to peritoneal dialysis (“PD”).
PD fluid called dialysate is provided typically in standard glucose levels. For example, in the United States, dialysate is provided typically in standardized glucose levels of 1.36%, 2.27% and 3.86% (corresponding to dextrose levels of 1.5%, 2.5% and 4.25%, respectively). The higher the glucose level, the higher the osmotic gradient caused by the dialysate, causing a larger amount of ultrafiltrate (“UF”) or waste water to be removed from the patient. The higher the glucose level, however, the more calories provided by the dialysate, and the more weight that can be gained potentially by the patient.
Patients accordingly typically use the lowest glucose level dialysate possible that will remove a needed amount of UF. Sometimes, however, the patient's therapy needs fall in between the standardized glucose levels of 1.36%, 2.27% and 3.86%. It may be desirable to use a blended glucose level of, for example, 2.0% glucose.
Tests can be performed on the patient to see how effective a particular dialysate is at removing waste and UF from the patient. For example, a peritoneal equilibrium test (“PET”) can be performed, which analyzes samples of dialysate taken after different dwell periods within the patient's peritoneum. The PET also requires analysis of the patient's blood. The PET is accordingly typically performed at a clinic.
In a clinical setting, it may be difficult if not impossible to blend dialysate solutions of different glucose levels to achieve a desired hybrid glucose level. A need therefore exists for a way to readily model hybrid glucose level dialysates.
The present disclosure provides an accurate and readily implemented method for modeling blended or hybrid glucose level dialysates. The method analyses each glucose level dialysate component separately and sums or integrates the results. This is done as opposed to actually blending the constituent glucose level dialysates, eliminating the need to manually or automatically blend the dialysates and preventing the possibility of error in blending. An error in blending for example leads to results that predict the patient's response to a glucose level blend that is different than what it is supposed to be.
The inventors have found, using mathematical models, e.g., via a three-pore kinetic modeling, that summing the results of individual dialysates having differing glucose levels yields an overall result that closely approximates the result of a blend of each of the components. For example, results for overall UF, urea Kt/V, cumulative creatinine removed, total carbohydrates (“CHO”) absorbed and total sodium removed showed very good correlation for two 1.5% plus two 2.5% dextrose level fills (corresponding to glucose levels of 1.36% of 2.27%, respectively) versus four 2.0% dextrose level fills. Tests performed and discussed in detail below also showed good correlation for patients with different peritoneal membrane transport types, including high, high average, low average and low patient types. The tests were performed using modeling simulated via a modified three-pore kinetic model discussed in more detail below.
It was also found that the model provides a way to estimate the UF and Kt/V based on a glucose concentration that is cumulative of multiple solution bags with different glucose concentrations, and which is independent of the order of infusion and glucose content of the solutions. It was further found that a linear relationship exists between (i) UF removed and % glucose and (ii) Kt/V and % glucose. Thus, one this linear relationship is learned for the patient, any desired final glucose concentration, not only the ones that could be obtained by mixing readily available solution bags, can be predicted.
It is accordingly an advantage of the present disclosure to provide a method of readily and accurately predicting the results for a dialysis treatment that uses a dialysate blended from different glucose level dialysates without having to actually blend such dialysates.
It is another advantage of the present disclosure to predict the results of a blended PD dialysate therapy for patients having different PD transport characteristics.
It is a further advantage of the present disclosure to develop linear relationships for UF removed and Kt/V versus glucose percentage, such that results for UF removed and Kt/V can be predicted for any final blended glucose level.
It is yet another advantage of the present disclosure to produce a database of the above linear equations, which are accessed using a two dimensional chart, wherein one dimension maps transport status as: high (“H”), high average (“HA”), low average (“LA”), and low (“L”) transport versus a second dimension which maps a particular therapy, e.g., eight hour, four exchange, two liter therapy.
The present disclosure addresses a growing need to know the peritoneal dialysis (“PD”) therapy outcomes when at least two different glucose concentrations are mixed to form a new blended solution, which is customized to suit the particular patient. The methodology discussed herein allows customized glucose results to be predicted, leading to a preferred mixture of standard solution, which can be mixed in real time by an automatic peritoneal dialysis (“APD”) machine.
The methodology uses relatively simple, linear equations that relate mixed final solution glucose concentration to therapy outcomes, such as net ultrafiltration (“UF”), weekly urea Kt/V, and creatinine clearance.
In one embodiment, a modified three-pore kinetic model of PD transport was used as the basis for the predictive mathematical model. One suitable modified three-pore kinetic model is described in Rippe B., Sterlin G., and Haraldsson B., Computer Simulations of Peritoneal Fluid Transport in CAPD, Kidney Int. 1991; 40: 315 to 325. Another suitable modified three-pore kinetic model is described in Vonesh E. F. and Rippe B., Net Fluid Adsorption Under Membrane Transport Models of Peritoneal Dialysis, Blood Purif. 1992; 10: 209 to 226, the entire contents of each of which are incorporated herein by reference and relied upon. Matlab™ (version 7.5.0.342, Mathworks Inc.) was used to construct the model.
The patient parameters used to illustrate the present method were obtained from data submitted to the assignee of the present disclosure in 1999 by centers around the United States and Canada participating in a national adequacy initiative program. The data were grouped in categories according to the patient's peritoneal transport status as: high (“H”), high average (“HA”), low average (“LA”), and low (“L”) transport statuses. A typical patient for each category was selected as shown in
For each Combination A to C, an 8-hour, 4-exchange therapy was analyzed using 2 liter fills for each unmixed or blended concentration. The difference in the simulated outcomes between unmixed and mixed solution conditions above have been summarized for key parameters such as net UF, urea Kt/V, creatinine clearance, glucose absorbed, and sodium removed. Detailed results are shown in the following Tables, 1A, 1B, 2A, 2B, 3A and 3B. Tables 1A, 1B, 2A, 2B, 3A and 3C show excellent correlation for each combination A to C, for each patient (identified in the tables as Patients 2, 5, 8 and 11) and thus for each patient clearance type H, HA, L, LA, respectively.
Table 1B for Patient 2 (high clearance) shows that when Combination A was modeled in two different combination orders of 1.5% versus 2.5% dextrose concentration combinations, the modeled results for mixed 2.0% concentration matched both unmixed combination orders very closely. Table 2B for Patient 8 (low clearance) shows that when combination B was modeled in two different combination orders of 1.5% versus 4.25% dextrose concentration combinations, the modeled results for mixed 2.88% concentration match both unmixed combination orders vary closely. Table 3B for Patient 5 (high average clearance) shows that when combination C was modeled in two different combination orders of 1.5% versus 4.5% dextrose concentration combinations, the modeled results for mixed 3.56% both unmixed combination orders vary closely.
Tables 1A, 1B, 2A, 2B, 3A and 3B
It should be understood that the equations shown in
The above simulations demonstrate that PD therapies conducted by infusing a mixture of readily available dextrose concentrations are equivalent to infusing each conventional solution one at a time as long as the cumulative dextrose concentration is the same. It is also shown that a linear relationship exists between the dextrose concentration and the outcomes.
The ability to demonstrate such equivalence is important in a variety of ways. First, the methodology provides a useful way to support regulatory claims in the area of solution mixing. Second, as seen in
Another key aspect of the present method is the generation of the linear simple relationships for UF and urea Kt/V, which allows any suitable glucose concentrations to be predicted.
It should be appreciated that the disclosed methodology and resulting apparatus can be extended to non-glucose based solutions, glucose-based solutions with lower sodium concentrations, or bi-modal solutions.
It should be understood that various changes and modifications to the presently preferred embodiments described herein will be apparent to those skilled in the art. Such changes and modifications can be made without departing from the spirit and scope of the present subject matter and without diminishing its intended advantages. It is therefore intended that such changes and modifications be covered by the appended claims.
