Heterogeneity and vulnerability of older patients

Abstract

Modern pharmacotherapy has to meet high-quality standards for the optimized treatment of every patient. On the one hand, standardization is indispensable for the determination of stable dosages and effect prediction and hereby is one of the major prerequisites for prescription safety and treatment efficacy. On the other hand, individualization of pharmacotherapy is mandatory because both patient-related and environmental factors lead to a wide variability of clinical phenotypes with critical relevance to drug treatment. Despite all standardization procedures and recommendations, pharmacotherapy is still characterized best as an individual experiment in daily practice; factors influencing the therapeutic response are never entirely known a priori, and the effect size cannot be securely calculated in advance. To minimize risks and maximize benefits, a careful estimation process for the risk-benefit ratio should take place at the beginning of every pharmacotherapeutic intervention. Today, information concerning the potential, and thus expected benefit, of a drug relies mainly on clinical studies in accordance with the principles of evidence-based medicine (EBM). Randomized controlled trials (RCTs) represent the data source with the highest quality, and general validity is claimed for results derived from such studies. However, significant limitations of RCTs are often overlooked; RCTs are preferably performed in populations that are artificially homogeneous by virtue of narrow inclusion and wide exclusion criteria. They thus do not actually represent the entire population with a given disease but may even only be typical for a minority of patients within this population. This strict patient selection mainly reflects methodological considerations and limitations in that clear drug effects may become diluted by confounders such as concomitant diseases or relevant conditions (e.g., kidney impairment). As the selected study population does not represent all patients with the disease, results should not be automatically generalized to the treatment of all patients with a given disease. Serious concerns about the so-called external validity of trial data should be triggered when important subgroups are severely underrepresented, as is still true for most trials almost systematically excluding the elderly (Bugeja et al. 1997; Lee et al. 2001; Dodd et al. 2011; Witham and McMurdo 2007).