‘Drugs Don't Work in Patients who Don't Take them’ (C. Everett Koop, MD, US Surgeon General, 1985)

This article refers to ‘The mortality risk of deferring optimal medical therapy in heart failure: a systematic comparison against norms for surgical consent and patient information leaflets’, by S. Zaman et al., published in this issue on pages 1401–1409. In both Europe and the United States, written informed consent is routinely obtained from a patient prior to surgery, other invasive procedures, or even some imaging studies. The text of the consent typically includes specific percentages for both the likelihood of mortality as a result of the intervention and common adverse events; a patient's signature indicates their willingness to proceed with the procedure and their understanding of the potential associated risks and benefits. Informed consent is expected to include a discussion of potential alternative therapies or interventions and the likely result of refusing the procedure or surgery.1 It has been suggested that written informed consent should also be obtained prior to the institution of medical therapies, but the process is less standardized in application.2 Patients are often informed about potential adverse effects of medications and are likewise told that the therapy will decrease mortality, reduce symptoms, or both; they are rarely presented with the type of precise statistics provided in written consents for surgical or invasive procedures. In this issue of the Journal, Zaman and colleagues have provided an analysis that clarifies the folly of our current approach, and questions why most clinicians have different thresholds to seek written informed consent in medical vs. more invasive therapies.3 These investigators performed a meta-analysis of randomized clinical trials of medical therapies in patients with heart failure and reduced ejection fraction (HFrEF) to determine the relative risk of 1-year mortality after the institution of angiotensin-converting enzyme inhibitors (ACEis), beta-blockers (BBs), and mineralocorticoid receptor antagonists (MRAs). Using these relative risks, they first calculated that a patient with HFrEF started on all three classes of HFrEF medical therapy would have a 90% 1-year survival, but only a 78% 1-year survival if all three therapies were delayed 1 year. They reasoned that the absolute excess annual mortality of deferring all three treatments for 1 year was an impressive 12 additional deaths per 100 patients. Next, the investigators surveyed clinicians in the UK to determine what absolute mortality would prompt informing a patient that a therapy was potentially associated with an increase in mortality. A full 85% of clinicians indicated their threshold to obtain written informed consent was a risk of 1 in 100, and 33% felt the boundary should be 1 in 1000. Thus, the risk of deferring all therapy for HFrEF for 1 year is 12 times higher than the risk that most clinicians consider to be the threshold for informed consent. Two other important points are made by these investigators. First, a delay of just 1 month for the three classes of medical therapies for HFrEF would result in one death per 100 patients. Second, deferral of an ACEi, BB, or MRA results in a yearly excess mortality of 4.4, 4.8, or 3 per 100 patients, respectively. Each of these risks, whether monthly for all therapies or yearly for individual therapies, crosses the threshold generally used to provide informed consent for invasive procedures. It is important to note that deferral of medical therapy comes in two forms—a lack of provider prescribing and patient non-adherence. We know that both are significant issues in heart failure patients. At first, one might reflect that few patients with HFrEF have all therapies deferred and that the current study is not relevant to current practice. However, Deschaseaux et al. used the Truven Health Market Scan database to determine that 42% of patients with a new diagnosis of heart failure were not prescribed any heart failure-specific therapy within 30 days of discharge.4 Unfortunately, the investigators were unable to determine how many of these patients had HFrEF compared with other forms of heart failure; the study needs to be repeated. We do know, however, that in general clinical practice the use of all three drug classes for HFrEF is not ideal. For example, Albert et al. noted that at the start of a quality improvement programme, ‘Get With The Guidelines’, just under one-third of all eligible HFrEF patients received MRAs at hospital discharge for an episode of acute heart failure.5 Moreover, despite the quality improvement programme, MRA use only increased from 28 to 34% over a 3-year period.5 Based on the yearly relative risk of mortality without MRAs from the current investigation, the low use of MRAs translates to an excess 3-year mortality of 9 patients per 100. Deferral of therapy for HFrEF can also be a result of patient non-adherence. Initial non-adherence to various medical therapies ranges from 7.4% with ACEis to 18.7% with proton pump inhibitors; at least 15% of HFrEF patients are non-adherent over a period of 3 months.6, 7 Poor medical adherence is not unexpected as heart failure patients have more complex medical regimens than patients with other common, chronic medical diseases.8, 9 An additional source of risk to the patient with HFrEF is the common use of less than optimal doses of guideline-directed medical therapies, documented in a number of registries and trials.10 The absolute risk of submaximal doses of drug therapies will be much harder to estimate than deferral of therapy, but is a threat nonetheless. What can be done to remedy our current illogical and unsafe approach to informed consent in patients with heart failure? It would be valuable for our clinical practice guidelines' writing committees to include rough estimates of the risk of delaying or stopping heart failure-specific medical therapies, especially as the committee members are reviewing the clinical trials used to inform the guideline recommendations. Additionally, new writing committees could be organized, charged with developing written informed consent templates for the initiation of medical therapy, rather than requiring every practitioner, specialist, and non-specialist to provide individual versions. If clinical practice guidelines for various chronic diseases provided these estimates, depicted in simple language, and/or graphically, both the provider and the patient could decide together about the most important medications to improve survival and which medications should be prioritized. In heart failure, as in other chronic diseases, the medications for the specific chronic disease make up less than one-third of all their medications; there is opportunity for some prioritization of medications.9 The study of Zaman et al. is a call to action to improve our goal of patient-centred care. We must re-examine our thresholds to seek written informed consent for medical therapy and strive for more consistent and granular informed consent for our patients. What C. Everett Koop, a former US Surgeon General, said in 1985 remains true today: ‘Drugs don't work in patients who don't take them’. Conflict of interest: none declared.