Disease progress models

Post, T. M., Freijer, J. I., Dejongh, J., Dan-hof, M. Disease system analysis basic disease progression models in degenerative disease. Pharm Res 2005, 22 1038-1049. 

A disease progression model describing the change in underlying disease during the course of the trial without treatment. 

Disease progression can be defined as the change in disease status over time. For the simulation of long-term administration of drugs intended to treat degenerative diseases, progression models can be very helpful and should be considered for model building and interpretation of the results. 

Especially for special therapeutic areas where it is unethical to treat patients for a longer time with a placebo, the development of disease progression models is of high interest because they allow a better discernment of the true treatment effect, for example, for these therapeutic areas Alzheimer s disease, Parkinson s disease, Osteoporosis, HIV, diabetes, and cancer. Furthermore, such models can help to differentiate whether the drug has only a symptomatic effect or a disease-modifying effect. The implementation of disease progression models also improves the reliability and acceptance of simulations. 

If the empirical approach is chosen, an empirical disease progression model can also be developed based on data collected during clinical trials or from public databases. A general empirical disease progression model has the following components ... 

Overall, the disease progression models allow a visualization of the time course of diseases under treated and untreated conditions and allow one to investigate in silico the impact of different therapeutic interventions. 

The linear disease progress model (Equation 20.2) assumes a constant rate of change of a biomarker or clinical outcome that reflects the disease status (S) at any time, t, from the initial observation of the patient — for example, at the time of entry into a clinical trial. The rate of change can be defined in terms of a baseline disease status (Sq) and a slope (a), which reflects the change from baseline status with time ... 

Finally a disease progress model can reflect more complex drug action. Phenomena such as a drug concentration-effect delay tolerance and rebound to both placebo and active treatments can be made using a linear offset model. These effects can be accounted for by including the appropriate terms. For instance a delay in onset can be accounted for by the addition of an effect compartment and tolerance and rebound... 

As with the offset model for the linear disease progress model, the effect of drug would be expected to disappear on cessation of therapy in this offset model. Again, a delay to the onset of drug effect can be incorporated with the use of an effect compartment component. 

The effects of a therapeutic agent (Ejp) on the progress of a disease may include both an immediate palliative effect and a reduction in the overall recovery time. Equation 20.11 describes the combination of these actions on the zero-asymptote disease progress model ... 

Figure 20.8 illustrates the four basic drug effect patterns when the input or output parameter changes with an exponential time course. As an example of this type of disease progress model, consider postmenopausal osteoporosis reflected by the net loss of bone mass after the menopause. Bone loss may be due to decreased formation or increased resorption of bone. Figure 20.9 illustrates the time course of bone mass change due to increased bone loss and the effect of administering a drug to reduce that loss. For example, raloxifene has been shown to be beneficial in women with postmenopausal osteoporosis (11). The pattern of increase in bone mineral density observed after treatment with raloxifene or placebo resembles the curves shown in Figure 20.10. However, the treatment duration in this dataset was too short to identify the actual mechanism of raloxifene effect on disease progress.

Within the past decade there has been significant interest in determining whether the use of clinical modeling and simulation software would increase the probability of conducting successful clinical trials (43). This approach incorporates the technique of pharmacokinetic-pharmacodynamic modeling that was discussed in Chapter 19 with the disease progression models described in Chapter 20. Although this type of a clinical development tool has considerable potential, the outcomes to date have been mixed. For example, this approach has identified the placebo-response rate for various disease states as an... 

TABLE 21.1 Example Database Format for Disease Progression Model... 

There are several models that need to be developed to describe the exposure-response surface. These include both PK and PD models, where the PD model includes models of disease progression and drug action. Prior to developing a model for disease progression, it is helpful to examine the different components of a disease progression model, and to understand the terminology associated with these models. 

The terminology presented here is consistent with terminology described and used previously (13, 14). In order to demonstrate the different types of drug action, example plots were generated using a simple linear disease progression model. 

The linear model assumes a constant rate of change of the disease status over time. The linear model can be dehned in terms of a baseline disease status ( o) and a slope parameter (a) and t is time after the initial observation of the disease. The equation for the linear model is given below and a NONMEM control stream implementing the linear disease progression model is provided in Table 21.2. 

When building a model for disease progression, it is often best to develop the disease progression model hrst, and then a model for drug effect is added. Typically, several disease progression models will be tested and the one that appears to best describe the time course for the markers of patient status is taken further to evaluate the addition of models for drug effect. 

TDAY is used here instead of time, which is usually in hours for the pk model. Using tday gives the disease progression model in days (or any unit of time such as weeks or months) rather than hours for several reasons. The first reason is that the parameter estimates are usually more... 

EST MAX=9990 SIG=3 NOABORT PRINT=1 METHOD=COND MSFO=base.msf The conditional estimation method is used here because the residual error for the disease progression model is additive. The conditional method with interaction can also be used as well. Because disease progression models can run for extended periods of time due to complex models and the large databases required, the use of the MSF file option is recommended. This option allows the job to be restarted if the minimization process is terminated for some reason (e.g., power failures). 

---