ERYTHROPOIETIC STIMULATING AGENT (ESA) DOSAGE DETERMINATION
An Erythropoietic Stimulating Agent (ESA) dosing system/method determines patient-specific ESA therapies for patients affected by insufficient hemoglobin production that may benefit from ESA treatment. The ESA dosing system includes a model that represents a process by which red blood cells are produced in humans. The model may include one or more parameters, the values of which are patient-specific. The model takes into account patient-specific historical hemoglobin (Hgb) data and corresponding historical ESA dosage data to estimate the patient-specific values of the model parameters, and determines a target therapeutic dose of the ESA that may maintain the patient's Hgb within a target range.
Latest MAYO FOUNDATION FOR MEDICAL EDUCATION AND RESEARCH Patents:
- Scalable multi-resolution electrode array for sensing and stimulating the brain
- Polypeptide-antibody complexes and uses thereof
- Methods and devices for electroporation for treatment of ventricular fibrillation
- Treating cardiovascular disease by selectively eliminating senescent cells
- Compositions and Methods of Treatment Using CRRL 191
The disclosure relates to modeling of biophysical parameters to determine pharmaceutical dosages.
BACKGROUNDAnemia causes an increased sense of fatigue, decreased stamina and exercise tolerance, fatigue, shortness of breath, decreased appetite, and decreased CNS functioning. Anemia can lead to the need for red blood cell (RBC) transfusions, with associated risks including bacterial and viral infections, volume overload, iron overload, and a variety of transfusion reactions.
Chronic Kidney Disease (CKD) and End Stage Renal Disease (ESRD) patients are at risk for anemia since RBC homeostasis requires normal kidney function. The kidneys play a critical role in erythropoiesis. Erythropoietic Stimulating Agents (ESAs) are used among these patients as a pharmacological replacement for the hormone erythropoietin (EPO), produced primarily by healthy kidneys, and to a small extent by the liver (Hepatic EPO). Other patient populations, including cancer patients, may also experience reduced levels of hemoglobin and may benefit from ESA therapy.
Hemoglobin (Hgb) values are a primary indicator of anemia. The Centers for Medicare & Medicaid Services (CMS) and National Kidney Foundation (NKF) have established the target range for Hgb values among ESRD patients to be between 10 g/dL and 12 g/dL. Hgb values below the desired minimum lead to an increased sense of fatigue and decreased stamina and are considered to be a risk factor for increased cardiovascular morbidity and mortality in ESRD patients. Patients with Hgb values under 10 g/dL suffer from the effects of anemia, including fatigue and reduced stamina and exercise tolerance, shortness of breath, decreased appetite and decreased CNS functioning, and reduced compliance. Anemia can lead to the need for red blood cell (RBC) transfusions, with associated risks including bacterial and viral infections, volume overload, iron overload, and a variety of transfusion reactions.
Hgb values above 12.0 g/dL are believed to create an increased risk of cardiovascular events such as stroke and myocardial infarction, cerbrovascular and cardiovascular mortality and morbidity. Patients with Hgb values over 12 g/dL are at risk of thrombosis, vascular access clotting (compromising effective dialysis therapy), hypertension, and increased risk of acute coronary syndromes or cerbrovascular accidents. These observations have led to the development of regulatory and quality standards which lead practitioners to try and maintain the hemoglobin values of ESRD patients within the narrow range of 10.0-12.0 g/dL.
In ESRD patients, as well as in other patient populations experiencing reduced hemoglobin levels, the biophysical system that regulates erythropoietin production does not function properly. ESAs are often prescribed to manage hemoglobin levels (anemia) in ESRD patients and in other patient populations. An ESA prescription may include, for example, intravenous injection of darbepoetin alfa (Aranesp®) or Recombinant Human Erythropoietin (rHuEPO). The current protocol for developing ESA prescriptions produces patterns of hemoglobin (Hgb) oscillation that subject patients to a cycle of overshoot and undershoot of target Hgb values. For example, when the patient exhibits a low Hgb, the dosage may be dramatically increased in an attempt to quickly raise Hgb levels. When the patient exhibits a high Hgb, interruption of ESA therapy (by greatly reducing the dose or withholding administration) may lead to under-dosing of the ESA, which, in turn, leads to an undershoot of Hgb values. The result is an undesirable fluctuation of Hgb levels above and below the target range. The period of the High-Low-High may take up to nine months for a complete cycle. Hgb values are often measured monthly, rendering Hgb cycling practically imperceptible.
In addition to the effects of low or high Hgb values upon ESRD patients, there are considerable administrative and financial impacts upon a dialysis facility if Hgb values are not maintained within the desired range. For example, the current protocol requires an ESA prescription to be developed one time per month per patient. Due to the cyclic variation in both Hgb levels and ESA dosage, considerable personnel time is used to review and adjust ESA dosage.
The current protocol also cannot project future actual ESA requirements, leading to difficulties with ESA inventory management. As a result of this uncertainty, dialysis facilities will often maintain large ESA inventories. However, because ESAs are relatively expensive, maintenance of large unused ESA inventories may not be financially optimal. In addition, Medicare and/or other insurance providers may impose penalties when patient Hgb levels exceed 12.0 g/dL for varying periods of time. These denials may occur retrospectively; that is, after the ESA has already been administered and the cost has been incurred by the dialysis facility.
In addition, patient Hgb values are often monitored monthly by regulatory agencies. A systematic pattern of high Hgb values can cause sanctions to be applied, which may include the creation of monitored compliance plans or even closure of a dialysis facility until a plan to achieve compliance is approved.
SUMMARYIn general, the disclosure describes system(s) and/or method(s) for determining Erythropoietic Stimulating Agent (ESA) dosing.
In one example, the disclosure is directed to a system comprising a biophysical simulation engine that represents a process by which red blood cells are produced in humans, wherein the biophysical simulation engine includes a plurality of parameters the values of which are patient-specific and a processing unit that receives patient-specific historical hemoglobin (Hgb) data and corresponding erythropoietic stimulating agent (ESA) dosage data, applies the Hgb data and the ESA dosage data to the biophysical simulation engine to estimate the patient-specific values of the plurality of parameters, and determines a target therapeutic dose of the ESA based on the patient-specific values of the plurality of parameters. The system may further identify, based on the target therapeutic dose, an equivalent dosing regimen titrated to available commercial doses. The parameters include one or more of an Erythroblast Production Rate, a Blast Mortality Fraction, a Reticulocyte Mortality Fraction, a Hepatic EPO, a RBC Average Lifespan, an EC50, and a setup EPO. Alternatively, the parameters include one or more of a Blast Forming Unit Input, a Colony Forming Unit Survival, a Recticulocyte Survival, an Erythropoietin Receptor Multiplier, a Red Blood Cell Lifespan, and an Erythropoietin Setup Rate. The processing unit may apply an optimization algorithm to estimate the patient-specific values of the plurality of parameters.
In another example, the disclosure is directed to a method comprising receiving patient-specific historical hemoglobin (Hgb) and corresponding patient-specific historical erythropoietic stimulating agent (ESA) dosage data, estimating patient-specific values for each of a plurality of parameters of a biophysical simulation model that represents a process by which red blood cells are produced in humans based on the patient-specific historical Hgb and corresponding patient-specific historical ESA dosage data, and determining a therapeutic dose that results in a predicted Hgb level stabilized within a target range based on the patient-specific values for each of the plurality of parameters.
In another example, the disclosure is directed to a method comprising receiving patient-specific historical hemoglobin (Hgb) and corresponding patient-specific historical erythropoietic stimulating agent (ESA) dosage data obtained during a descriptive period, estimating patient-specific values for each of a plurality of parameters of a biophysical simulation model that represents a process by which red blood cells are produced in humans based on the patient-specific historical Hgb and corresponding patient-specific historical ESA dosage data, determining patient-specific simulated Hgb values for a prescriptive period based on the patient-specific values, and identifying a target therapeutic dose of the ESA that maintains the patient-specific simulated Hgb values in a target range during the prescriptive phase.
The details of one or more examples are set forth in the accompanying drawings and the description below. Other features and advantages will be apparent from the description and drawings, and from the claims.
The disclosure generally relates to systems and/or methods that design patient-specific Erythropoietic Stimulating Agent (ESA) dosing regimens. The ESA dosing system and/or methods described herein may result in determination of patient-specific ESA dosing that achieves and sustains adequate Hgb values for patients receiving ESA therapy.
The ESA dosing techniques described herein may be used to determine patient-specific ESA dosing for any available ESA therapy. These ESAs may include, but are not limited to, Erythropoietin; Epoetin alpha (Procrit®, Epogen®, Eprex®); Epoetin beta; darbepoetin alpha (Aranesp®); Methoxy polyethylene glycol-epoetin beta; Dynepo; Shanpoeitin; Zyrop; Betapoietin; and others.
In addition, the ESA dosing techniques described herein may also be applicable to a wide variety of patient populations, including, for example, End Stage Renal Disease (ESRD) patients, Chronic Kidney Disease (CKD) patients, cancer therapy patients, or any other patient population having insufficient hemoglobin production that may benefit from ESA treatment such as anemia secondary to HIV infection. In addition, the ESA dosing techniques described herein may also be applicable to multiple modes of ESA therapy delivery, including intravenous (IV) delivery, subcutaneous delivery, oral delivery, biopump, implantable drug delivery devices, etc.
In recent years there has been much controversy regarding optimal Hgb target values in ESRD patients, as well as controversy over the impact of swings in Hgb values. Based on currently available clinical data, there is widespread agreement that the target value for Hgb should be somewhere between 10 and 12 grams/deciliter (g/dL), and that the probable desired optimal range is 11.0-12.0 g/dL. There is also growing agreement that stable Hgb values are more conducive to patient well-being than are wide oscillations in Hgb values. The system described herein enables care providers to identify dosing regimens that will establish and maintain Hgb values in the target range for the majority of their patients.
The system includes a patient-specific biophysical simulation model that, based on a patient's historical response to ESA therapies, determine a target dosing level which can be translated to a dosing regimen titrated to available commercial doses. The dosing regimen thus obtained can be configured to simultaneously achieve and sustain adequate and stable Hgb values for extended periods of time as well as minimize or eliminate Hgb oscillations (commonly known as Hgb cycling). The total amount (and cost) of ESA administered may also be reduced or minimized. If the patient's overall medical condition remains stable, Hgb values have been shown, using the techniques described herein, to remain stable at a given target level. If the patient's underlying medical condition changes, the system includes a diagnostic system'which can be used to establish a new target dosing level that may restore Hgb values to a desired target level in a minimum of time.
In some examples, the system/method creates a recommended intravenous (IV) ESA dosing regimen including a dose level and dose administration frequency. Care providers can tailor the frequency of ESA administration enabling effective and efficient use of supporting staff time.
The purpose of system 10 is to help care providers develop ESA dosing strategies that avoid creating the oscillations in Hgb values for patients that are characteristically created by existing protocols, and that provide stabilized Hgb levels within a target Hgb range.
Patients with ESRD have a deficiency of the hormone erythropoietin (the endogenous ESA), and, as a result, they are severely anemic. Anemia (hemoglobin<10.0 g/dL) is a risk factor for mortality in ESRD patients, and patients with anemia have poorer quality of life than non-anemic patients. Patients receiving an ESA also have an increased risk of cardiovascular events (stroke, myocardial infarction) if their hemoglobin rises above 12.0 g/dL. These observations have led to the development of regulatory and quality standards which lead practitioners to try and maintain the hemoglobin values of ESRD patients within the range of 10.0-12.0 g/dL. In addition, other patient populations may also receive ESA therapy, including CKD patients, cancer therapy patients, and other patients who would benefit from ESA therapy, and it shall be understood that ESA dosing system 10 may also be applicable to these and other patient populations. Thus, although some portions of this description may refer specifically to ESRD or CKD patients, it shall be understood that ESA dosing system 10 and the techniques implemented therein may also be applicable to other patient populations.
Patient-specific responses to ESA therapy are dependent upon a variety of factors, including total body iron storage status, extracellular volume fluid status, inflammation, residual kidney function, hemorrhage, and variations in the dose effectiveness of ESA among ESRD patients.
The majority of ESRD patients who need ESA therapy are currently receiving one of two formulations: rHuEPO, FDA approved for the treatment of anemia in patients with chronic renal failure in 1985, or darbepoetin alfa (Aranesp®), FDA approved in 2001. rHuEPO has an average half life of five to seven hours, requiring frequent administration. darbepoetin alfa was designed to have a longer half-life of 25-27 hours. The example simulation engine was designed to monitor patient response to darbepoetin alfa. However, adjustments to the simulation engine may allow for similar simulations to be conducted for patients receiving o rHuEPO, in transition from rHuEPO to darbepoetin alfa, or other ESA therapies.
Due its longer half life, darbepoetin alfa requires approximately five days for complete elimination. This allows providers to administer the drug less frequently. But the extended half-life of darbepoetin alfa, in combination with red blood cell dynamics, contributes to a confounding physiological consequence. After an administration of darbepoetin alfa, RBC production is enhanced for up to 26 days. This delay, if not factored into the design of the prescription, sets up Hgb oscillation. It is not uncommon for patients to experience 12-18 months of Hgb “overshoot” and “undershoot” as providers try to reestablish an adequate and stable Hgb level following existing protocols.
The system uses an operational approach that includes all the factors that generate patient Hgb values. The system and associated processes and models described herein may help providers design ESA therapies that eliminate Hgb oscillations and achieve adequate and stable Hgb values within target levels.
Due to its longer half life, darbepoetin alfa requires approximately five days for complete elimination from the serum, and has a prolonged period of pharmacological activity. This allows providers to administer the drug less frequently. But the extended half-life of darbepoetin alfa, in combination with red blood cell dynamics, creates a physiological consequence. After an administration of darbepoetin alfa, RBC production is enhanced for up to 26 days. This delay, if not factored into the design of the prescription, sets up Hgb cycling. It is not uncommon for patients to experience 12-18 months of Hgb “overshoot” and “undershoot” as providers try to establish an adequate and stable Hgb level following existing protocols. The system accounts for feedback and delay in the erythropoietic process by establishing a target dosing level, assisting with the design of a dosing regimen, and monitoring results over time.
In current practice it is often desirable to reduce the frequency of administration in order to capture reduced administrative costs. When a patient is switched from bi-weekly to monthly dosing, for example, current practice is to double the dose and then seek the optimal regimen using current protocols. This introduces a potential round of Hgb cycling. However, it has been found using the presently described biosimulation techniques that reducing the frequency of administration by a factor of two requires far more than twice the previous dose. If such a decrease in frequency is otherwise desirable, the system permits the identification of the required dose to sustain adequate Hgb values. In other examples, the system may determine that increasing the frequency of administration with optimal doses may, in spite of increased administrative costs, reduce total cost due to significantly reduced amount of the drug required.
The presently described biosimulation techniques utilize dynamic modeling. Dynamic modeling is a framework, consisting of a language and a set of concepts. These are embedded in a process for representing, understanding, explaining and improving how dynamic systems (erythropoiesis, for example) work, how they perform over time, and how they respond to inputs (such as ESA administration).
There are several commercial packages available to build and simulate dynamic models, including iThink®, available from Isee Systems, Inc.; Stella®, available from Isee Systems, Inc.; Vensim®, available from Ventana Systems, Inc.; Powersim Studio 8, available from Powersim Software AS; Berkeley Madonna™, developed by Robert Macey and George Oster of the University of California at Berkeley; and other commercially available software packages. The example described herein was implemented using iThink version 9.3 and the examples provided herein are described using iThink syntax and conventions. It shall be understood, however, that the specific implementation described herein is one example of how the biosimulation model may be implemented, and that equivalent models may be constructed in each of the aforementioned commercial packages, in other commercially available packages, in customized software packages, or in application specific software programs and/or systems, and that the disclosure is not limited in this respect.
The model manages the dynamic linkage that exists between the pharmacokinetics and pharmacodynamics of the ESA in question with the dynamics of the RBC chain. Further, the model may be embedded in a data processing system that enables effective ESRD anemia management both at the individual and group level.
The process of constructing the example model was to consult with subject matter experts to learn how selected variables are related and then to translate those relationships to a specific model in the chosen syntax. Various decisions are made in the model building process concerning levels of aggregation/disaggregation required to achieve the model's purpose. In other examples, levels of aggregation/disaggregation may be modified to achieve the same purpose, while exposing differing biophysical behaviors over time.
The model building process includes decisions about boundaries of the model that are consistent with the model purpose. This is referred to as establishing the extensity of the model. Other examples of the model may include revisions to the model's extensity as described below.
The specific example of the model described with respect to
Typically in scientific studies of the factors that relate to a so-called dependent variable, one performs various studies of correlation, analysis of variance, principal components, etc. Parameters of a Dynamic Model, however, are identified and used differently than in statistical studies. Once the boundary of a dynamic model is defined, the parameters describe exogenous inputs to the model. The parameters of a dynamic model describe operational variables (as, in general, do endogenous model variables as well) in that they describe causal factors of the behavior being simulated. As an example, the parameter Erythroblast Production Rate (EPR) is one parameter to the biosimulation model. This parameter describes the rate at which erythroblasts are created per day. The value of this parameter is not merely correlated to the RBC count, but, all other parameters equal, a given value for EPR will cause a certain number of RBC's to exist. Note that parameters may be constants or complex mathematical expressions, representing aggregates of external subsystems. Note that parameters selected and defined for a dynamic model depend upon the definition of the model's boundary.
The system develops a targeted dosing plan, and then anticipates changes in the patient's response to the ESA therapy, which allows for an equally targeted response that reduces or avoids Hgb oscillations. The result may be that more ESA therapy patients have hemoglobin values maintained within the desired range of 10.0-12.0 g/dL, or within 11.0-12.0 g/dL.
The modeling process takes into account not only endogenous physiological factors that regulate red blood cell (hemoglobin) values, but also the patient in order to achieve and maintain an adequate, stable Hgb level. This model is unique for each individual patient, and this allows for inclusion of patient-specific components of anemia management.
Table 1 lists the seven patient-specific parameters utilized in this example, provides a definition of each, and sets forth example default minimums and maximums used in this example of the model. Alternative examples may use a different parameter set, yet still describe the dynamics of Hgb response to various dosing regimens.
The Erythroblast Production Rate (EPR) represents the number of erythroblasts created per day within the bone marrow, but outside the boundary of this example of the model.
The leakage flow named blast mortality (602 in
The Reticulocyte Mortality Fraction (RMF) is similar in effect to the BMF. In this example of the model, darbepoetin alfa stimulates RBC production by increasing the survival rate of reticulocytes in the marrow.
Hepatic EPO is shown in
RBC Average Lifespan is included as a parameter for this example of the model. It is known that, while in healthy individuals, the RBC Lifespan is about 120 days, for ESRD patients on dialysis, the RBC Lifespan is shorter in duration. RBC Average Lifespan is named Average Lifetime (606) in
In practice, the EC50 of an agent is the concentration that produces a response half way between the baseline and maximum response for a given time period. Usually a measure of potency, this parameter is used differently in this example of the model as a measure of what is known in the field as EPO resistance. Various medical conditions such as inflammation, infection, and the presence of ESA antibodies can decrease an individual patient's response the ESA therapy. This example of the model uses a single measure of EPO resistance. Alternative examples may use separate values for each cause of EPO resistance which could potentially produce an improved simulation.
As explained above, the RBC chain is represented in this example of the model as an array of 12 so-called bins of RBC cells. In the initial phase of a simulation, named the Setup Phase, a steady state RBC count is established in each of the 12 bins, corresponding to the initial actual Hgb value for an individual patient in the second, Descriptive Phase of the simulation. (The Descriptive Phase is described below.) To establish a steady state RBC count, this example of the model is provided with what is mathematically equivalent to an externally administered dose.
The user of the model, or the software system into which the model is integrated, develops an initial estimate of patient-specific parameter values, as described below. Initial estimates of parameter values may be manually adjusted using the interface to the biophysical simulation engine.
A dynamic model, as expressed using the selected commercially available simulation package or any of the others listed above, is defined by its degree of aggregation among selected variables, model boundaries, exogenous parameter values, time period to be simulated, and time increment (DT) to use for the simulation. A model so expressed simulates proposed causal relationships among its elements, as distinct from correlated relationships. As such, a model so defined represents a theory of dynamic behavior of a system that can be tested in a laboratory, confirmed, and refined. The dynamic modeling process often includes simulation and testing of a proposed dynamic hypothesis using a specific model, testing and validation of the dynamic hypothesis, followed by revisions to any of the model elements to improve model performance. Dynamic modeling is an iterative process in which the dynamics of the system are represented, understood, and explained in order to improve the simulation of the dynamic system under study, in this case erythropoiesis for iron replete ESRD patients on dialysis receiving darbepoetin alfa. The scope of the claims presented below shall include all the iterates of the example models described herein and those which may evolve in future examples.
Referring again to the example core model shown in
Table 2 lists the correspondence between the seven patient-specific parameters listed in Table 1 and the variables shown in
The following are illustrative equations for the example model shown in
The simulation performed using the equations described above results in a numerical approximation of the solution to a set of differential equations that describe accumulations in the chosen stocks (which represent integrals) as determined by their respective inflows and outflows (which represent derivatives). Specifically, the user of this model supplies, for an individual patient, historical Hgb values, historical darbepoetin alfa doses, the time period to be simulated, and the time increment, DT. The simulation, embedded in a simplified optimization routine (described below) then enables the user to determine a target dosing level and an associated dosing regimen that will obtain the desired Hgb values as long as the patient's current medical condition remains relatively unchanged.
In one example, the biophysical simulation model may employ an adaptation of the Monte Carlo method to estimate parameter values. However, it shall be understood that other non-linear optimization routines may also be used, and that the disclosure is not limited in this respect.
In this example, 100 or fewer simulations of the Monte Carlo method may be to choose an optimum value. However, in alternative examples of the model, thousands or tens of thousands of simulations might be run in a reasonable amount of time, which may allow for a more complete assessment of the distributions of each of the seven parameters. Further, this example of the model may not provide a unique solution. The same proposed dosing regimen might be developed for one patient with a low EPR, BMF, and RMF as for a patient with a high EPR, BMF, and RMF. Alternative examples of the model may allow potential classification of patients of the first or second type. In practice, however, proposed dosing regimens, though non-unique, may be quite adequate, resulting in 60% to 90% or more of the patients at a DCF achieving and sustaining Hgb values within the target range.
Each simulation is executed in three phases: Setup, Descriptive, and Prescriptive. Each phase is defined over a specific number of days. As described below, the Setup Phase extends from Day −200 to Day 1, the Descriptive Phase extends from the day number associated with a patient's first historical Hgb value (chosen by an analyst or chosen automatically by an automated software system) to the most recently available Hgb value or administered darbepoetin alfa dose. The Prescriptive Phase extends from the simulation day number of the first potential dosing date (generally one week after the end of the Descriptive Phase) to a simulation day number at which a proposed dosing regimen, given the selected parameter values, produces stabilized simulated Hgb values at the desired target value. The extent of each phase is identical for the two modes of simulation: the Monte Carlo Mode that generates a collection of random parameter values and associated simulated Hgb values, and the single simulation mode, in which proposed dosing regimens are developed or revised.
For the individual patient, the Descriptive Phase is used to select a set of parameter values that produces simulated Hgb values that match the historically observed Hgb values in response to historical drug doses over the duration of the Descriptive Phase. There will be a combined set of parameter values (e.g. the baseline progenitor input, baseline progenitor mortality in the absence of darbepoetin alfa, the EC50 of darbepoetin alfa, the level of protection provided, and the lifespan of circulating RBCs). This phase serves to define the ‘pathophysiological state’ of the patient, and the patient's sensitivity to the drug.
The setup phase begins at ‘−200 days’, i.e. prior to ‘Day 1’, the day at which historical data is available (the beginning of the Descriptive Phase). The purpose of the setup phase is to identify a set of parameters that bring the system into a steady state (flat-line Hgb level equal to the patient's first Hgb value) prior to Day 1, and then simultaneously enable the system to follow the patient's response to darbepoetin alfa during the Descriptive Phase. The parameter Setup EPO (see Table 1) is used primarily in the Setup Phase to represent a mathematical dose of darbepoetin alfa, which, together with other parameter values in play at during the Setup Phase, achieves the results described immediately above. The parameter Setup EPO has no effect in subsequent phases but is rather replaced by either historical doses (in the Descriptive Phase) or proposed doses (in the Prescriptive Phase).
Given an appropriate set of parameter estimates, simulated Hgb values will respond to historical doses during the Descriptive Phase by generating simulated Hgb values that approximate the waxing and waning of historical Hgb values. The Partial Monte Carlo method, complemented by additional manual adjustments and/or automated adjustments, if required, is used to identify the best fit described above. Further, the best fit is defined as the simulated Hgb values within the extent of the Descriptive Phase, selected from a collection of simulations, which have a mean square error with respect to actual Hgb values in the Descriptive Phase of approximately 0.25 g/dL.
During the Prescriptive Phase, one or more of the following steps may be performed with fixed parameter values. A therapeutic dose (which may be a per session therapeutic dose or a weekly therapeutic dose (WTD) depending, among other things, upon the particular ESA to be prescribed) may be determined. In addition, in the event of a WTD, an analyst or automated system may introduce ‘sample’ dosage regimens, searching for a dosing regimen (dose and frequency) that delivers the equivalent of the WTD at a minimum cost. Frequently dose “pulses” may be required to quickly elevate Hgb values or avoid a projected undershoot. The dosage regimen is refined to bring the patient quickly and smoothly within the target Hgb range and to sustain that value. The selected dosing regimen may be extended several months into the future and remains effective as long as the patient's underlying medical condition remains relatively stable. In alternative examples, the search for a WTD and the selected dosing regimen may be implemented using automated software algorithms.
As described above, in this example of the biophysical simulation model, parameter values are optimized across the Setup and Descriptive phases using non-linear optimization methods (such as Monte Carlo techniques). However, it shall be understood that the present disclosure is not limited in this respect. Alternative examples may include, for example, other optimization strategies such as simplex algorithms and maximum likelihood estimators or other non-linear computational algorithms known to those of skill in the art.
The pharmacokinetic (PK) section of the model (
It is recognized that additional pathophysiological subsystems (and comorbid conditions) may influence Hgb levels over time. Alternative examples of the model may include, for example, iron availability and treatment regimens, bleeding, and the influence of inflammation on EPO resistance.
Other examples of the model may include additional parameter refinements, supporting disaggregation of biophysical subsystems, when these variations improve fulfillment of the purpose of the model.
ESRD patients on dialysis frequently suffer blood loss for various reasons, such as bleeding from the access point to the patient's bloodstream or gastrointestinal bleeding. Other patient populations may also experience hemorrhages for various reasons. This example of the model includes the ability to specify up to four periods during which the patient experiences a hemorrhage.
Although it is possible to identify individual factors that influence Hgb values over time, it is the interaction between these factors that influence the time course and magnitude of Hgb values in response to darbepoetin alfa and other ESAs. The present ESA dosing system includes a mathematical model that provides definitions of individual components, and then allows the controller to conduct simulations that reliably predict the behavior of the system as a result of perturbations that may be systematically introduced by an analyst or automatically by an automated software system.
The ‘Weekly Therapeutic Dose’ (WTD) is a theoretical value that defines the weekly dose that will ultimately maintain the patient's Hgb values at the midpoint of the target range. WTD is determined by systematically varying a fixed weekly dose and observing Hgb concentrations during the prescriptive phase of the simulation. The therapeutic dose determined by the ESA dosing system may be a per session therapeutic dose (PSTD) or a weekly therapeutic dose (WTD) depending, among other things, upon the particular ESA to be prescribed. In the case of Aranesp®, the system determines a WTD and then determines a dosing regimen that will deliver the equivalent of the WTD. In the case of other ESAs, the therapeutic dose determined by the system may be equivalent to the actual dosing regimen recommended for the patient.
The WTD derived in the Prescriptive Phase of the simulation is the clinician's guide to developing a dosing regimen. A dosing level and the frequency at which to administer the chosen doses may then be chosen from commercially available doses. For some ESAs, the WTD may be unequal to commercially available doses (in the case of Aranesp®, for example), and this may require a mix of doses be applied at various frequencies that together will deliver a dose equivalent to the WTD. Further, on the date the dosing regimen is to be started, the patient may currently be either above or below the target range, with either an upward or downward trend in Hgb values. In such cases “pulse” doses must be found that will quickly and smoothly achieve Hgb values within the target range.
(1001) denotes the Setup Phase
(1002) denotes the Descriptive Phase
(1003) denotes the Prescriptive Phase
(1004) portrays simulated Hgb values in each phase
(1005) reports historical ESA doses administered in the Descriptive Phase
(1006) reports historical Hgb values measured during the Descriptive Phase
(1007) reports three pulse doses to initiate the Prescriptive Phase, designed to arrest the concurrent downward trend in Hgb values
(1008) reports the proposed dosing regimen that will sustain Hgb values at 11.5 g/dL
(1009) is the simulation day number of the first dose in the three dose pulse (1007)
(1010) is the amount of the dose to be administered as Dose A
(1011) is the interval in days for Dose A
(1012) is the end date for Dose A
(1013)-(1016) are the analogues of Dose A specifications, to be applied as Dose B
(1017)-(1020) are the analogues of Dose A specifications, to be applied as Dose C
Thus, in this case, the steady state dosing regimen is found to be: “Starting on day 125, give three weekly doses of 25 mcg, followed by alternating doses of 25 mcg and 40 mcg every 21 days”.
One concept of dynamic modeling in the example system described herein is that recommended strategies (i.e., dosing regimens) are hypotheses as opposed to “black box answers”. These hypotheses are to be tested by follow up measurements of actual future Hgb values of the patient and either confirmed or rejected. Both confirmation and rejection of a hypothesis provide insight and understanding of how the process under study actually operates, which is a major goal to be achieved from a dynamic modeling perspective, generally and particularly. To that end, the system may also include, for example, components and tools for follow up, analysis, learning, revision, improved anemia management skills, and ultimately the well being of the patient.
The example system described is a clinically applicable set of tools designed to address and resolve Hgb cycling. The example system includes of a collection of components that have been loosely coupled by means of various software components. Alternative examples of the system may include tightly integrated modules of functionality, providing an information system that supports anemia management for individual patients receiving therapy at one or more Dialysis Care Facilities (DCF).
The purpose of the system is to capture, cleanse, maintain, transform, analyze, and create data and information required by clinicians to effectively manage anemia concerns for a population of individual patients.
The first step in treating a population of patients is to select applicable patients. In one example, the system is designed to treat iron replete patients who have a minimum of 6 recorded Hgb values. Patient data described above is assembled and organized for processing and maintenance. In other examples, the system may include iron metabolism components, potentially enabling the inclusion of patients who were not iron replete during the period in which Hgb values were obtained.
Obtaining and maintaining individualized simulation parameters, whether those described above, or refinements and parameter improvements, are stored in a database. Alternative examples may include effective classifications of Hgb response profiles, potentially creating improved methods of treatment.
Individualized recommended prescriptions are stored in a database which permits overall analysis of ESA consumption and improved management of associated costs. In this example of the system, the database is implemented in a commercially available spreadsheet program. Alternative examples may include implementation of customized database application using commercially available database engines, data transformation tools, analysis and reporting.
Recommended prescriptions are reviewed and approved by authorized providers. The example system provides recommend dosing regimens. In addition, clinicians' anemia management skills may be improved through use of the system.
Approved prescriptions may be entered into a provider's order management system. The ESA dosing system may be loosely coupled with a medical order management system. Alternative examples may include software components that tightly integrate each step in the process from a data and information perspective. Alternative examples may also provide improved ESA consumption management which may significantly reduce a variety of operational costs, such as reduced drug consumption, reduced carrying inventories, reduced spoilage, reduced administrative costs, and reduced administrative costs associated with preventable emergent medical issues.
The ESA dosing system may also include data collection tools designed to monitor the compliance of drug administration with medical orders. Hgb cycling often arises from dose misadministration. Using these tools, the ESA dosing system may detect dose misadministration and prompt as well as design corrective interventions.
As described above, the recommended dosing regimen, once approved and ordered, is a hypothesis awaiting confirmation or rejection, each of which improves insight. The ESA dosing system may include, for example, weekly monitoring of Hgb values for a minimum of 12 weeks, allowing the clinician to detect and diagnose the causes of observed deviations of actual Hgb values from those that were predicted by the simulation.
The BOT chart is a tool used by clinicians in the implementation of the methodology supported by the system.
-
- 1101 administered iron (Venofer®)
- 1102 Mean Corpuscular Volume (MCV)
- 1103 simulated Hgb values for a portion of the Descriptive period
- 1104 actual Hgb values
- 1105 projected Hgb values in response to planned therapy
- 1106 hospitalized 7 days for pneumonia
- 1107 transferrin saturation and iron values (resting upon one another, but with different scales)
- 1108 boxes represent recommended and approved dosing regimen
- 1109 dots within boxes represent actual aranesp doses administered
- 1110 open box represents recommended and approved dose either missed or not yet administered
- 1111 upper bound of target Hgb range (13.0 g/dL)
- 1112 upper bound of optimal target Hgb range (12.0 g/dL)
- 1113 lower bound of optimal target Hgb range (11.0 g/dL)
- 1114 lower bound of target Hgb range (10.0 g/dL)
A given patient's BOT, containing these and other data (such as the time course of vital signs) enables the clinician to develop a comprehensive picture of the patient's overall condition. Axis labels and scales are not shown in
The process may also include a structured change control process. The ESA dosing system may be designed to anticipate changes or replacements to any or all of the system components. For example, the ESA dosing system may anticipate changes in an individual patient's underlying medical condition. These changes may require resimulation of a new recommended prescription, or searching for a new set of parameter values and then developing a new recommended prescription. The ESA dosing system may include tools whereby an analyst can retrieve previously modeled patients and begin anew. Alternative examples may include information system components that maintain histories of identified parameter values, prescriptions, and changes over time of the patient's medical condition. This information may improve insights and an operational understanding of the relationship between the progression of CKD for ESRD patients on dialysis and Hgb response profiles.
The graph 154 of
The model uses a so called pre-descriptive period to establish an erythropoietic equilibrium with an RBC count, which reflects the Hgb level that is near the first observed Hgb result in the descriptive period. In this example, the pre-descriptive period in the model is 201 days in duration, running from day −200 to day 0. This is the period of time the body requires to establish equilibrium in the presence of a theoretical (mathematically applied) daily ESA dose. The model uses the parameter values displayed on the left of the
The search for parameter values stops when the Mean Square Error (MSE) between the simulated Hgb values and observed Hgb values in the descriptive period is sufficiently small. In
The simulation based approach solves the problem of overshoot and undershoot by associating the post administration exponential decay in ESA concentration levels with the delays involved in red blood cell production. The model accounts for the production and mortality of RBC precursor cells. By providing estimates of RBCs “in the pipeline”, the provider can extract advice from the biophysical simulation engine that will create a dosing plan that will achieve an adequate and stable Hgb value within a target range.
As discussed above, an output of the process is the calculation of the weekly therapeutic dose (WTD) required for a stable Hgb at the target value or within a target range. The WTD is the theoretical weekly dose required to achieve the intended therapeutic response. That is, a dose level which, if administered weekly, would stabilize a patient's Hgb at the target level.
The simulation of the prescriptive phase of the patient included a weekly dose titrated to deliver the equivalent of 25 mcg of darbepoetin alfa per week. As shown by the simulated Hgb levels for the prescriptive period, Hgb would stabilize at 11.5 g/dL after a little over 60 days.
Once the system determines the optimized WTD, the system may assist providers in finding the most effective combinations of available dosing levels at the optimal frequency of administration that will deliver the required WTD, and as a result, achieve and maintain the desired Hgb value. This may include, for example, titrating available dosing levels that will deliver the equivalent of the WTD. The example illustrated in
For other patient simulations, WTD values that are not equal to available unit doses require experimentation to determine the optimal dosing strategy. For these patients, the process may guide the provider to a titration scheme that achieves the intended result.
After a proposed prescription is approved and administration has begun, the patient's situation will invariably change. A hospitalization, an infection that increases resistance to ESA therapy, or a hemorrhage may occur that changes the patient's response to ESA therapy.
In accordance with another aspect of the example ESA dosing system, it has been determined that Hgb measurements taken in the prescriptive period that differ from the projected (simulated) response are reliable indicators that the value of at least one of the patient-specific parameters has changed. Identification of changes in observed Hgb levels from the projected response may be used to prompt focused assessments about changes in the patient's condition that may lead to effective corrections.
For example, variance in Hgb levels from the projected response may be related to a condition or situation that was not present during the descriptive phase. Re-modeling may be used at this point to seek an alternative set of patient-specific parameters values. The newly updated parameters may then be used to yield an effective corrective action, that is, an updated WTD, to restore an adequate and stable Hgb value.
Monitoring of Hgb levels during the prescriptive period may therefore be a part of the systemic solution. This part of the process is a probe that scans for changes in the patient's condition, develops corrective actions, and communicates the required changes in a timely and effective manner.
The system determines a therapeutic dose that results in stabilized Hgb within the target range (206). In the event that the therapeutic dose is a WTD, the system may identify one or more dosing regimens that deliver the equivalent of the WTD (208). The one or more equivalent dosing regimens may be developed for a 30, 60, 90 or other day regimen, up to six months, for example. Other variables for the equivalent dosing regimens may include the dosage given per dialysis session and/or the number and frequency of doses. Multiple dosing regimens from among available titrations of ESA therapy may be identified, and a dosing regimen that minimizes one or more variables such as the dosage given per dialysis session, the number or frequency of doses, cost, or other appropriate factors.
In addition, corrective therapies may be identified, diagnostics may be ordered, statistics summarized and group performance reports developed.
If applicable, the patient-specific parameter values may be manually or automatically adjusted to account for known hemorrhages (236) or transfusions (238). The system then determines the therapeutic dose based on the optimized patient-specific parameter values that may maintain the patient's Hgb within a target range (240).
In one example, “optimal” patient-specific parameter values are identified through a type of Monte Carlo simulation that minimizes the mean square error between simulated Hgb values and the patient's actual Hgb lab history. A Monte Carlo simulation is a method of randomly selecting model parameter values to be used in a simulation in order to seek optimal values selected from the results of a large number of simulations.
Simplified (e.g., less processing intensive) versions of the Monte Carlo simulation may run only 100 or so simulations, whereas more robust versions may allow thousands or tens of thousands of simulations. Expanding the Monte Carlo sample space by orders of magnitude may improve the reliability of the proposed prescriptions and/or reduce the need for expert judgment. In addition, as described herein, other non-linear optimization routines may also be used to obtain the patient-specific parameter values, and the disclosure is not limited in this respect.
The example biophysical simulation engine described above was limited to those ESRD patients that were iron replete throughout their descriptive periods. This permits exclusion of iron metabolism components from the simulation engine. However, it shall be understood that iron metabolism components may be added to the example system to accommodate patients who experience periods of iron deficiency, and possible reduced responsiveness to ESA therapy.
As described above, the biophysical simulation engine estimates Hgb values based on the red blood cell count. An alternative would be to add model components that include whole blood hematocrit calculations. For example, blood plasma and fluid dynamics model components could be added. These alternatives may account for any hemodilution and hemo-concentration effects. Once the hematocrit is known, an estimate of the Hgb level could be derived that would be more accurate than the current estimate.
As discussed above, variances from projected Hgb levels may be used as a diagnostic, enabling users to anticipate and potentially prevent undesirable outcomes. In each case where significant variances were observed, a model based upon the patient's new reality may be reconstructed, allowing the system to create a revised prescription and to continue pursuing an adequate and stable Hgb level with a revised therapy.
In addition to improvements in the management of anemia, with adequate and stable Hgb values achieved, the patient will have more stamina to comply with the rigors of life on dialysis. As a result, hospitalizations may decrease, missed sessions may decrease, mortality may be reduced, and dietary restrictions may be more valued and observed. The patient may obtain the presence of mind to effectively engage with managing the details of creating their own health outcomes. Because missed ESA doses perturb Hgb values, the information provided by this process greatly reduces the risk of missed doses going uncorrected.
Elimination of Hgb cycling for an individual patient eliminates a number of patient health risks. Hgb cycling is an indicator that all the systems of the body are experiencing alternating periods of excessive and diminished oxygen supply. It is believed that this variation in oxygenation leads to increased hospitalization rates and mortality. Stable Hgb values improve patient quality of life and reduce overall health risks. In addition, it may make it easier for health care providers to recognize the onset of new comorbidities in their patients who have a stable Hgb while on therapy since in general new medical problems may lead to a fall in Hgb.
In one example, the ESA dosing system is applied to patients receiving dialysis and darbepoetin alfa (or other ESA) therapy. CKD patients who are not on dialysis, however, also may require ESA therapy. CKD is a progressive disease and generally leads to the initiation of renal replacement therapy, most frequently, dialysis. Hgb cycling among CKD patients not on dialysis has also been observed. Thus, in other examples, the system may also be applied to those CKD patients currently not on dialysis, improving their overall health and stamina, and designing therapies to postpone the initiation of dialysis and the associated rigors and costs.
In other examples, the ESA dosing system may be applied to any patient requiring ESA therapy.
In general, ESRD, CDK, and other patients with adequate and stable Hgb values are easier for care providers to manage. In addition, the ESA dosing system may create a new perspective for care providers concerning RBC homeostasis. Insights gained by the successful management of Hgb values using the system and methods described herein may be, at least partially, transferable to patients who have not been assessed using the ESA dosing system described herein. For example, such insights may permit more accurate and optimal dosing regimens to be designed. Thus the ESA dosing system may reduce the complexity, time, and cost of caring for patients and improve effectiveness at the same time.
The ESA dosing system and methods described herein may also improve management of ESA inventories at Dialysis Care Facilities. Using this system, accurate projections of ESA requirements may be made, reducing excess inventory. Because ESA are relatively expensive, this reduction in inventory may result in great savings per year in inventory costs. Projected, precise dosing regimens for each patient receiving dialysis at a Dialysis Care Facility (DCF) for the future (90 days, for example) equips the DCF with a more accurate estimation of the required ESA inventory levels. This can reduce the cost of waste and other costs associated with carrying excessive inventories.
The system and method described herein allows creation of dosing regimens that achieve adequate and stable Hgb values that also consume a minimum amount of ESA drugs. Retrospective assessments of the data has produced an estimate that ESA costs may be reduced by as much as 46% or more.
The system may include an analysis and reporting subsystem that provides, for example, “at a glance” overviews of patients with below target Hgb values, in range, or above target Hgb values. Maintaining adequate and stable Hgb values for a higher percentage of patients may enable providers to spend less time per patient, and allocate more time to the care of patients with emergent medical issues.
The ESA dosing system may be used to increase the efficacy and efficiency of administered ESAs while achieving adequate and stable Hgb values. This is a primary concern of CMS and may represent significant cost savings.
The ESA dosing system may provide a proven evidence-based assessment of the effectiveness of (or inadequacy of) various dosing regimens or protocols. By means of the ESA dosing system described herein, providers are equipped with a target dosing level heretofore unknown. Other examples of the ESA dosing system may include dosing regimen quality metrics that give providers the data they need to continuously improve their anemia management practices.
National and private insurers are moving for pay for performance reimbursement policies. Other examples of the ESA dosing system may include tools to assist with an objective performance measurement system for the management of anemia.
By resulting in more stable Hgb levels, the ESA dosing system described herein may also decrease the amount of un-reimbursed ESA that has been administered. In sum, the derived therapies may continuously improve patient outcomes, financial performance for providers, and multidisciplinary care team effectiveness.
Experience has shown that providers using rHuEpo are more able to achieve target Hgb values than are providers using darbepoetin alfa. However, rHuEpo alfa may be administered up to three times per week, whereas darbepoetin alfa may be administered weekly, bi-weekly, or even monthly. Providers have attempted to switch to the use of darbepoetin alfa in order to reduce operating costs only to decide at a later time to revert back to darbepoetin alfa due to uncontrolled Hgb cycling. Other examples of the ESA dosing system may include tools to assist providers in transitioning from epoetin alfa to darbepoetin alfa and simultaneously maintaining adequate and stable Hgb values.
Recombinant human erythropoietin (rHuEPO, Epogen, EPO) has a shorter half-life than darbepoetin alfa and is therefore easier to administer. However, dialysis providers utilizing rHuEPO as an ESA must manage and administer rHuEPO at each dialysis session, resulting in increased operational costs, increased risk of infection, and higher turnover on their ESA inventories. Although the techniques are described herein with respect to darbepoetin alfa, it shall be understood that the techniques could be adapted to any form of ESA. The ESA dosing system may be used to assist dialysis providers worldwide in making a successful transition from rHuEPO to darbepoetin alfa and secure the benefits of achieving adequate and stable Hgb values along with the reduced operational costs of less frequent ESA administration.
The successful utilization of the modeling described herein has repercussions that may extend beyond the use of ESAs in patients. For example, the systems, methods, and techniques described herein may be extended to the administration of other drugs with prolonged half-life or extended release formulation. The pharmokinetic studies that are required by the FDA do not provide renal or hepatic function. Likewise, genetically determined differences in drug metabolism are not evident to clinicians until an adverse effect of under- or over-dosing of the drug is noted. The application of this methodology to drug administration may allow faster determination and use of optimal drug dosages, and highlight individual patient differences in the clearance and metabolism of drugs. The current trial and error method of drug administration requires improvement if we are to more safely administer drugs in a patient population with increasing incidence of kidney and liver disease, and increasing utilization of drugs with longer half-lives.
The example ESA dosing model described with respect to
In addition, the ESA dosing techniques described herein may also be applicable to a wide variety of patient populations, including, for example, ESRD patients, CDK patients, cancer therapy patients, HIV patients, or any other patient population having insufficient hemoglobin production that may benefit from ESA treatment. In addition, the ESA dosing techniques described herein may also be applicable to multiple modes of ESA therapy delivery, including intravenous (IV) delivery, subcutaneous delivery, oral delivery, biopump, implantable drug delivery devices, etc.
Model inputs 1222 represent the patient historical hemoglobin data 1226 and the patient historical ESA dosage data to be entered into the model. In this example, the historical ESA data may include either historical Aranesp data 1228 or historical Epogen data 1230, depending upon the ESA therapy used by the particular patient. Other ESA dosage data may also be entered, and the disclosure is not limited in this respect.
Fe (iron) status indicators 1224 represent the patient historical iron data 1232 and/or the patient historical transferrin saturation data, if any, to be entered into the model. This allows the user to take the patient's iron levels into account when running the ESA dosage simulation. Patients who are iron deficient, indicated in part be a transferrin saturation value below 20 percent, may not have enough iron in the blood to combine with mature reticulocytes produced by the bone marrow to create a sufficient number of mature red blood cells containing hemoglobin. If the simulation is not able to fit the historical data, the Fe status indicators 1224 may help the user to better interpret the patient's clinical status and define corrective therapies.
Hgb low 1218 and Hgb high 1220 represent the low and high values of the desired hemoglobin range. For example, the Centers for Medicare & Medicaid Services (CMS) and National Kidney Foundation (NKF) have established the target range for Hgb values among ESRD patients to be between 10 g/dL and 12 g/dL. These or other hemoglobin values appropriate for the patient or the patient's condition may be entered as the low and high hemoglobin values, respectively.
As described above, the biophysical simulation may employ an adaptation of the Monte Carlo method to estimate patient-specific parameter values. It shall be understood that other optimization routines may be employed, and the disclosure is not limited in this respect.
Blast Forming Unit Input (BFU INPUT): the number of erythroid burst forming units entering the erythropoietic process each day.
Colony Forming Unit Survival (CFU SURV): the fraction of colony forming units that survive apoptosis in the absence of an ESA.
Reticulocyte Survival (RETIC SURV): the fraction of reticulocytes that survive reticulocyte atrophy, which may be caused by a deficiency in hemoglobin building blocks such as iron, folate, or vitamin B12, among others.
Erythropoietin Receptor (EpoR) Multiplier (EPOR MULT): the value by which Kd is amplified to generate a response within the developing RBC cell a strong enough reaction to prevent apoptosis.
Red Blood Cell Lifespan (RBC LIFESPAN): average lifespan (in days) of a red blood cell.
Erythropoietin Setup Rate (EPO SETUP RATE): a mathematical value applied during the setup period that raises the simulated hemoglobin to a level equal to the observed hemoglobin on the first day of the descriptive period.
For each of the six parameters shown in
If the Monte Carlo switch for a particular parameter is turned on, the simulation will obtain the value of that parameter for each run of the Monte Carlo simulation via a random number generator as described below.
For each of the six parameters shown in
For each parameter, MC 1285, 1295, 1305, 1315, 1325, and 1335, represents a function that generates a random number between the values of MIN and MAX for each run of the Monte Carlo simulation. For example, BFU INPUT MC 1285 represents a function that generates a random number between the values of BFU INPUT MIN 1281 and BFU INPUT MAX 1283 for each run of the Monte Carlo simulation.
BFU INPUT 1280, CFU SURV 1290, RETIC SURV 1300, EPOR MULT 1310, RBC LIFESPAN 1320 and EPO SETUP RATE 1330 are either the previously determined parameter value obtained from a Monte Carlo run or user selected parameter values (input, for example, via sliders shown in
BFU INPUT CALC 1284, CFU SURV CALC 1294, RETIC SURV CALC 1304, EPOR MULT CALC 1314, RBC LIFESPAN CALC 1324, EPO SETUP RATE CALC 1334 are the values obtained from the best fit run of a Monte Carlo simulation. Once the best fit run is determined, the parameter values determined from that best fit run may be used to determine a therapeutic dose that may be administered in the prescriptive period. A therapeutic dose of an ESA is that dose which causes a patient's hemoglobin values to achieve and sustain the target hemoglobin value as long as the patient's clinical condition remains stable. However, the user may want to attempt to improve upon the initially obtained best fit results by manually selecting a value for one or more of the parameters (such as via the sliders shown in
Suggested BFU minimum (Sugg BFU min) 1286 and Suggested BFU maximum (Sugg BFR max) 1288, Sugg CFU surv min 1296 and Sugg CFU surv max 1298, sugg retic surv min 1306 and sugg retic surv max 1308, sugg EPOR mult min 1316 and sugg EPOR mult max 1318, sugg RBC LIFE 1326 and sugg RBC LIFE 1328, and sugg EPO setup min 1336 and sugg EPO setup min and sugg EPO setup max 1340 may be the results obtained from a Monte Carlo run. Should the user choose to perform a subsequent Monte Carlo run, these values may be the suggested values to use for the respective minimum and maximum values to use as lower and upper bounds from which random values will be drawn in the subsequent Monte Carlo run.
To determine the MSE for each run, the total number of hemoglobin values in the descriptive period is counted (1341, 1344). The squared difference of the simulated hemoglobin 1348 and the patient specific historical Hgb data 1226 is determined (1350). The squared differences are summed over all days of the descriptive period (1352, 1354). The MSE 1356 is the sum of the squares divided by the total number of hemoglobin values in the descriptive period 1344 (displayed in this example in box 1342). The MSE for each run is determined, and the run with the lowest MSE is determined to be the “best fit run.”
EPO SETUP RATE 1330 is one of the patient-specific parameters to be determined during the simulation. CLINICAL ARANESP DATA 1228 is the patient-specific historical Aranesp dosage data during the descriptive period. In this example, a recommended dosing regimen for Aranesp is divided into three separate doses to be administered on given days, dose A 1370, dose B 1371, and dose C 1372, to deliver the equivalent of the weekly therapeutic dose (WTD) determined by the simulation. Dose A START 1368 is the earliest day on which dose A may be given, as well as the day on which Dose A is to commence. ARANESP INPUT (micrograms) 1376 is a the total dose on a given day of the simulation from all sources: EPO setup rate 1330, Clinical Aranesp Data 1364, and Rx Protocol 1374. Rx Protocol 1374 can be of two types, it is either the sum of Aranesp dose A 1370, Aranesp dose B 1371, and Aranesp dose C 1372 or, it is Aranesp WTD 1373, depending on the value of WTD Switch 1371. Aranesp WTD 1373 is equal to the value of Example 2 Weekly Therapeutic Aranesp Dose Amount 1377, which is a user-entered dose, to be applied in the simulation beginning on First Prescriptive Sim Day Number 1212 (see
A published parameter referred to as the “ESA Kd” is stored by the ESA dosing system and applied in the model. The ESA Kd is a known value for each ESA that may be entered by a user and stored by the system. Kd refers to the dissociation constant of the ESA being used and the erythropoietin receptor (EPOR).
In this example, the ESA Kd values shown in
The calculated value “EPOR fraction bound” refers to the percentage of eporeceptors on the surface of BFU-E cells that have bound the ESA being used for therapy. Once a minimum percentage is reached, the rate of programmed cell death (apoptosis) is decreased by means of reactions within the cell in response to bound eporeceptors on the surface of the cell.
For example, if the EPOR fraction bound 1510 is greater than a minimum percentage (such as 10%) then the apoptosis rate decreases. Alternatively, if the EPOR fraction bound 1510 is less than the minimum percentage, the apoptosis rate increases. The effect of the EPOR fraction bound on the apoptosis rate is described in more detail below with respect to
CFU/E chain 1536 represents cell replication that occurs through a number of generations. The number of generations considered by the model may vary in a specified range; for example, from 12 to 31 generations. BFU INPUT 1280 is the number of blast forming units committed to forming red blood cells which is input into model at 1534. Due to cellular division, the input of each successive generation in the CFU/E chain 1536 is twice the output of the previous generation. A replication interval 1538 describes the length of time required for each generation to replicate. For example, the replication interval may be in the range of 0.75 to 1.25 days, or other appropriate interval. CFU/E cells are the cells within the erythrocyte lineage that are subject to apoptosis and responsive to apoptosis-sparing ESA therapy. The total number of cells within CFU/E chain 1536 is reduced by apoptosis 1543, as moderated by the ESA concentration. CFU Survival 1290 is a patient-specific parameter defined as the fraction of colony forming units that survive apoptosis in the absence of an ESA. EPOR fraction bound 1510 and EPOR multiplier 1310 determine the fractional amount of apoptosis sparing 1541 in the presence of an ESA. This lowers the rate of apoptosis 1543, increasing the number of dividing cells 1540 that survive apoptosis and go on to become erythroblasts 1542, and eventually, erythrocytes.
Erythroblast/Reticulocyte development chain 1544 represents maturation of the number of erythroblasts that survived apoptosis 1542, 1544. A maturation time 1546 describes the length of time required for each generation of erythroblasts to mature. For example, maturation time may be in the range of 3 to 5 days, or other appropriate maturation time. The total number of reticulocytes leaving development chain 1544 is reduced by reticulocyte atrophy 1548. Reticulocyte atrophy 1548 is determined by the patient-specific parameter Reticulocyte Survival 1300, defined as the fraction of reticulocytes successfully mature in the presence of required complementary materials, such as iron, folate, and B12, for example. Reticulocyte survival is also influenced by infection and inflammation. In the case of infection, bacteria compete with the maturing cells for iron, reducing the iron available to form hemoglobin in the maturing cell. In the case of inflammation, available iron is sequestered, effectively reducing iron availability to the maturing cell. In the current embodiment of ESA Dosing System 1240, assessments on the status of complementary materials, inflammation, and infection are made by an expert model user. The ESA dosing model may include software algorithms to simulate the status of complementary materials, inflammation, and infection in regard to reticulocyte survival, and the disclosure is not limited in this respect. The number of maturing cells that survive reticulocyte atrophy is the number of reticulocytes 1550 produced by the bone marrow.
The total cells in circulation 1566 (reticulocytes in circulation+RBCs in circulation) together with the volume of distribution 1472 gives the simulated hematocrit 1568. The hematocrit multiplied by a constant gives the simulated hemoglobin 1570.
The graph 1620 of
As mentioned above, although the examples were presented herein with respect to Aranesp and Epogen, it shall be understood that the ESA dosing techniques described herein may also be applicable to other types of ESA therapies, other patient populations and alternative routes of administration. In general, to apply the ESA dosing model to other ESA therapies, the ESA-specific constants ESA half life and ESA Kd would be entered into the ESA dosing model. For example, the ESA half life for the particular ESA would be taken into account in the model at the same point as the Aranesp half life 1418 in
The following are illustrative equations for the example model shown in
The techniques described in this disclosure, including functions performed by a processor, controller, control unit, or control system, may be implemented within one or more of a general purpose microprocessor, digital signal processor (DSP), application specific integrated circuit (ASIC), field programmable gate array (FPGA), programmable logic devices (PLDs), or other equivalent logic devices. Accordingly, the terms “processor” “processing unit” or “controller,” as used herein, may refer to any one or more of the foregoing structures or any other structure suitable for implementation of the techniques described herein.
The various components illustrated herein may be realized by any suitable combination of hardware, software, or firmware. In the figures, various components are depicted as separate units or modules. However, all or several of the various components described with reference to these figures may be integrated into combined units or modules within common hardware, firmware, and/or software. Accordingly, the representation of features as components, units, or modules is intended to highlight particular functional features for ease of illustration, and does not necessarily require realization of such features by separate hardware, firmware, or software components. In some cases, various units may be implemented as programmable processes performed by one or more processors or controllers.
Any features described herein as modules, devices, or components may be implemented together in an integrated logic device or separately as discrete but interoperable logic devices. In various aspects, such components may be formed at least in part as one or more integrated circuit devices, which may be referred to collectively as an integrated circuit device, such as an integrated circuit chip or chipset. Such circuitry may be provided in a single integrated circuit chip device or in multiple, interoperable integrated circuit chip devices, and may be used in any of a variety of pharmaceutical applications and devices.
If implemented in part by software, the techniques may be realized at least in part by a computer-readable data storage medium comprising code with instructions that, when executed by one or more processors or controllers, performs one or more of the methods described in this disclosure. The computer-readable storage medium may form part of a computer program product, which may include packaging materials. The computer-readable medium may comprise random access memory (RAM) such as synchronous dynamic random access memory (SDRAM), read-only memory (ROM), non-volatile random access memory (NVRAM), electrically erasable programmable read-only memory (EEPROM), embedded dynamic random access memory (eDRAM), static random access memory (SRAM), flash memory, magnetic or optical data storage media. Any software that is utilized may be executed by one or more processors, such as one or more DSP's, general purpose microprocessors, ASIC's, FPGA's, or other equivalent integrated or discrete logic circuitry.
Various examples have been described. These and other examples are within the scope of the following claims.
Claims
1. A system, comprising:
- a biophysical simulation engine that represents a process by which red blood cells are produced in humans, wherein the biophysical simulation engine includes a plurality of parameters the values of which are patient-specific; and
- a processing unit that receives patient-specific historical hemoglobin (Hgb) data and corresponding historical erythropoietic stimulating agent (ESA) dosage data, applies the Hgb data and the ESA dosage data to the biophysical simulation engine to estimate the patient-specific values of the plurality of parameters, and determines a therapeutic dose of the ESA based on the patient-specific values of the plurality of parameters.
2. The system of claim 1, wherein the processing unit further identifies, based on the therapeutic dose, an equivalent dosing regimen titrated to available commercial doses.
3. The system of claim 1, wherein the parameters include one or more of an Erythroblast Production Rate, a Blast Mortality Fraction, a Reticulocyte Mortality Fraction, a Hepatic EPO, a RBC Avg Lifespan, an EC50, and a setup EPO.
4. The system of claim 1, wherein the parameters include one or more of a Blast Forming Unit Input, a Colony Forming Unit Survival, a Recticulocyte Survival, an Erythropoietin Receptor Multiplier, a Red Blood Cell Lifespan, and an Erythropoietin Setup Rate.
5. The system of claim 1, wherein the processing unit applies an optimization algorithm to estimate the patient-specific values of the plurality of parameters.
6. The system of claim 1, further comprising a user interface that displays a graph of the patient-specific historical hemoglobin (Hgb) data and the corresponding historical erythropoietic stimulating agent (ESA) dosage data during a descriptive phase.
7. The system of claim 1, wherein the user interface further displays a graph of patient-specific simulated hemoglobin data and the therapeutic dose of the ESA during a prescriptive phase.
8. A method comprising:
- receiving patient-specific historical hemoglobin (Hgb) and corresponding patient-specific historical erythropoietic stimulating agent (ESA) dosage data;
- estimating patient-specific values for each of a plurality of parameters of a biophysical simulation model that represents a process by which red blood cells are produced in humans based on the patient-specific historical Hgb and corresponding patient-specific historical ESA dosage data; and
- determining a therapeutic dose that results in a predicted Hgb level stabilized within a target range based on the patient-specific values for each of the plurality of parameters.
9. The method of claim 8, wherein estimating patient-specific values for each of the plurality of parameters comprises applying Monte Carlo methods to estimate the patient specific values.
10. The method of claim 8, further comprising identifying one or more dosing regimens that deliver the equivalent of the therapeutic dose.
11. The method of claim 8, further comprising:
- receiving patient response data concerning actual Hgb levels of the patient under the identified dosing regimen;
- comparing the actual Hgb levels of the patient under the identified dosing regimen to the predicted, stabilized Hgb level;
- identifying variations from the predicted, stabilized Hgb level; and
- re-estimating the patient-specific values for each of the plurality of parameters of the biophysical simulation model to obtain an updated therapeutic dose based on the identified variations.
12. The method of claim 8, wherein determining the therapeutic dose comprises:
- receiving a target Hgb level; and
- iteratively applying the patient-specific values for each of the parameters to a series of proposed therapeutic dosages until the target Hgb is obtained.
13. A method comprising:
- receiving patient-specific historical hemoglobin (Hgb) and corresponding patient-specific historical erythropoietic stimulating agent (ESA) dosage data obtained during a descriptive period;
- estimating patient-specific values for each of a plurality of parameters of a model that represents a process by which red blood cells are produced in humans based on the patient-specific historical Hgb and corresponding patient-specific historical ESA dosage data;
- simulating patient-specific Hgb values for a prescriptive period based on the patient-specific values; and
- identifying a therapeutic dose of the ESA that maintains the patient-specific simulated Hgb values in a target range during the prescriptive phase.
14. The system of claim 1, wherein the parameters include one or more of a Blast Forming Unit Input, a Colony Forming Unit Survival, a Recticulocyte Survival, an Erythropoietin Receptor Multiplier, a Red Blood Cell Lifespan, and an Erythropoietin Setup Rate.
15. A method comprising:
- receiving patient-specific historical hemoglobin (Hgb) and corresponding patient-specific historical erythropoietic stimulating agent (ESA) dosage data;
- optimizing patient-specific values for each of a plurality of parameters of a model that represents a process by which red blood cells are produced in humans to determine a best fit with the patient-specific historical Hgb and corresponding patient-specific historical ESA dosage data; and
- determining a therapeutic dose of the ESA that maintains a patient-specific simulated Hgb values in a target range.
Type: Application
Filed: Jan 4, 2011
Publication Date: Jul 25, 2013
Applicant: MAYO FOUNDATION FOR MEDICAL EDUCATION AND RESEARCH (Rochester, MN)
Inventors: Craig L. Hocum (Winona, MN), James T. McCarthy (Rochester, MN), David P. Steensma (Wellesley, MA), David Dingli (Rochester, MN), James L. Rogers (Rochester, MN), Edward J. Gallaher (Hillsboro, OR)
Application Number: 13/519,843
International Classification: G06F 19/12 (20060101);
