Therapy optimization

Therapy optimization

Patients with a given disease may differ dramatically in the desired and undesired effects of one and the same standard drug therapy. Extreme examples are:

  • total non-responsiveness and
  • severe adverse effects

reported for virtually all drugs in smaller or larger patient subsets. These phenomena are driven by hidden or unconsidered heterogeneity among patients and within the disease entity itself. Easiest to avoid are undesired effects caused by inconsideration of therapy-relevant, clinically manifest patient conditions such as comorbidities or pregnancy. The more hidden heterogeneity is gradually comprehended and dealt with using a fast-growing array of disease and patient markers. Marker discovery is largely driven by the molecular analyses of patient subgroups with distinct treatment responses identified in clinical studies. This deserves emphasis as clinical studies are frequently but incorrectly thought to reduce treatment individuality.

Validated patient and disease heterogeneity markers are applied to maximize outcomes and to reduce toxicities and costs of drug therapies. Specifically, they help to:

  • select and later adjust drug and dosage
  • verify treatment adherence
  • avoid or early detect undesired drug effects.

Along the way, disease markers gradually transform the traditional organ- and histology-based disease classification into one more relevant for treatment decisions. Therapies are also optimized by the consideration of the patient’s treatment objectives and drug preferences.

Drugs and patients in need of

All therapies require optimization. The individual intensity and urgency is determined by the pharmacological and toxicological drug properties, by the course of the disease, and by patient-specific factors. Drugs requiring optimization and monitoring are usually rich in serious undesired effects and/or they are safe and effective only in a narrow concentration range. Patient-specific factors include his/her therapeutic objectives and drug preferences, contraindications such as age, pregnancy, or comorbidities, and the individual, usually hardly predictable profile of undesired drug effects. Responsiveness prediction based on (tumor) genotyping is used mainly in oncology and it is partly driven by high costs of newer drugs.  

Tools

Therapy optimization utilizes both surrogate (laboratory) markers and clinical symptoms constituting a contraindication or suggestive of an undesired drug effect. Therapy-optimizing markers are technologically indistinguishable from those applicable to diagnosis or prediction of progression and prognosis. A specific exception are concentration measurements of active or toxic drug forms. From the viewpoint of pharmacology, therapy optimizing markers interrogate either the pharmacokinetics or pharmacodynamics of a drug. Predictive markers are deployed prior to, assessing markers during or following treatments. Predictive markers are based on averages of many patients and they are therefore less accurate than assessment markers - in an individual patient, the interrogated pharmacokinetic or pharmacodynamics effects may be modified by additional factors specific to this patient.  

Pharmacodynamics markers are more informative for the desired or undesired drug effects than pharmacokinetic ones. This is due to the more immediate relatedness between clinical effects and pharmacodynamics than with pharmacokinetics (drug application -> pharmacokinetics -> pharmacodynamics -> clinical effects). For this reason, once established, pharmacodynamics markers tend to be more widely used than pharmacokinetics markers. However, pharmacodynamics markers are much more difficult to develop since they frequently require invasive sampling of diseased tissue. In consequence, most pharmacodynamics markers interrogate processes taking place (e.g. coagulation tests during anticoagulant therapy) or measurable (e.g. transaminases during drug hepatotoxicity) in the easily accessible blood compartment. If affected by functionally relevant genetic variability, a drug’s pharmacodynamics can be also interrogated by genotyping of blood-derived DNA, as can pharmacokinetics.

Therapy-optimizing: classification, tools, and examples

Informative for Prediction Assessment
To be deployed Prior to therapy During or after therapy
Applications

Selection

-drug

-dosage

Avoidance of adverse drug events

Selection

-drug

-dosage

Detection of adverse drug events

Verification of therapy compliance

Examples

Pharmacokinetics

-GFR estimation (e.g. metformin)

-TPMT genotyping (6-mercaptopurine)

Pharmacodynamics

-EGFR genotyping (erlotinib)

-HER2 immunohistochemistry (trastuzumab)

-echocardiography (anthracyclines)

Pharmacokinetics

-GFR estimation (e.g. metformin)

drug plasma concentration (e.g. cyclosporin A)

Pharmacodynamics

-Thrombocytosis (heparins)

-INR (warfarin & phenprocoumon)

liver transaminases (e.g. methotrexate)

-echocardiography (anthracyclines)