About

A 70yo retired nurse was referred with renal failure, migraine, reflux, asthma and anxiety, with systolic blood pressures (BPs) up to 190 treated with telmisartan 40 mg and hydrochlorothiazide 12.5 mg daily. Her regimen was broadened, keeping to the most effective antihypertensive drugs published in the large randomised clinical trials, but at ‘very low’ doses, nebivolol, 1.25 mg, and perindopril 2.5 mg + indapamide 0.625 mg alternate days (if dose-response is not plotted logarithmically, these doses appear in fact nearer mid-range). After 9 years, her weight was down 6kg, she had no cardiovascular symptoms and her systolic BPs remained under 150.

The enormous range of useful drugs now available, along with demand by ill and symptomatic patients, and much drug marketing, has made polypharmacy – multiple concurrent drug therapies – difficult to resist and rationalise. Polypharmacy is more prevalent with age as comorbidities accrue. Active symptomatic disease merits effective drug treatment but in each patient, prescribers are challenged by which drugs are essential, particularly for ‘prevention’.

Drug treatments, in spite of all of the potential benefits and our best intentions, contribute to significant numbers of hospitalisations and deaths. There is much socially, culturally and historically which would appear against drug treatments. There are many warnings in media, by clinicians and packaged with drugs about potential harms. A priority for clinicians and patients is ‘first do no harm’.
Most patients prefer not to take drugs, perhaps because of common knowledge that all drugs are to some extent poisons and may be deleterious.

Persistent symptoms and disease, along with enthusiasm and anxiety to get ahead of unfolding disease can prompt inadvertent overtreatment. A little treatment, if effective, may be better, sometimes much better, than none. Once a drug proves effective, dose becomes a useful ‘lever’ with which to establish maximum possible benefit with acceptable risk of harms.

Placebo-controlled RCTs focus on efficacy and usually only address one drug at a time. Additive or synergistic efficacy and toxicities with drug combinations are problematic to study and have been little addressed. Early antihypertensive trials used combinations and established the importance of the addition of a diuretic, but no combination was ever shown to lower systolic BP by more than 20 mm Hg.

Mathematical modelling of effects at different doses suffers from the limitations of the available data and the natural variability of disease and physiology between patients. Precise dosing can be titrated in acute and curable disease on the basis of the patient’s evolving symptoms, physical signs, and relevant tests and imaging. However, most disease can only be controlled, not eliminated.

When lower drug dose proves clinically sufficient, obvious advantages are the greater tolerability and safety, along with the less use of manufacture consumables and costs, i.e. ‘less can be more’. Dosing is problematic in the treatment of persistent illness, attempting to anticipate the uncertainties of disease trajectory and the long time periods required to titrate dose against any clinical response.

Doses in preventive treatment in asymptomatic patients are necessarily difficult to monitor and a default has increasingly become fixed doses which are based on high, sometimes very high, doses employed in published clinical trials which often recruit less complicated patients.

We have reviewed clinical experience and the largest clinical outcomes trials to better define drug dose response. Because of large ‘NNTB’ (number needed to treat for one patient to benefit) in prevention, typically 100-200, highest useful dose in each indication requires careful analysis to weigh possible benefit against multiple potential harms. Unlike in active disease, in prevention drugs are taken continuously for long periods so that benefits accrue with time, along with a range of adverse effects. Data on safety, especially mortality, are necessarily limited to the amount of
investment in clinical trials. Ascertainment of the optimal dose for the individual patient is problematic and a default has often been the maximum tolerated dose, which may only be appropriate in malignancy. There appears no other human endeavour where dose is maximised (e.g. soil science and enrichment; even the most enticing kitchen ingredients when used to excess can impede the result!)

Management of much illness suffers from limitations of continuity and follow-through. Any new insights we present emerged around much practical feedback from patients on long term drug treatments re possible adverse events, tolerability and wellness, particularly pertinent in preventive cardiology. There is now an increasing volume of clinical trials data, much of which is publicly available and we can only benefit from increasing independent drug review and governance but safety is difficult to reconcile, especially with newer products.

It is important we calibrate doses employed across different settings and bring colleagues and hospitals up to speed with optimal prescribing and dosing. We have focussed on the largest studies and some of our interpretation of the data may initially appear counter-intuitive. Benefit must exceed harms, especially in long term prescribing, best achieved by accurate diagnosis and drug targeting. For the individual patient, harms overtake benefits as dose is increased, creating essentially a ‘J-curve’. It has recently become apparent that the maximum useful daily  maintenance dose of aspirin is substantially less than the 75-100 mg daily popularly prescribed, increasing persistent unintended, unfavourable effects long term. Safety demandsavoidance of excess dose.

Guidelines are only a guide and may be influenced by biases related to drug funding, marketing and even geopolitical priorities. A supported and trusted international reference group which can assimilate all of the relevant evidence might have much utility, perhaps culminating in an accessible library of data and e.g. accepted ED50s for different drugs and indications. Personalised prescribing should obviously allow for the size of the patient, the severity of their illness and any impediments to
drug elimination, in particular renal failure.

It is meanwhile hoped that this website can provide a useful guide for pharmacists, physicians and patients in the pursuit of optimal prescribing.

Simon B Dimmitt FRACP FCSANZ
Clinical Professor, University of Western Australia
Conjoint Professor, University of Newcastle