Physicians struggle every day to pick the right drug dosage for the treatment and prevention of disease. For the acute illnesses, efficacy is evident within hours or days. For the prevention of chronic disease, however, the outcome is uncertain at best. Therefore, we rely on randomized clinical trials to provide evidence that a specific drug and dosage are safe and effective.
Unfortunately, because of the limited average follow-up of 3-5 years, randomized clinical trials (RCTs) do not provide efficacy and safety information for lifetime therapy that is often advocated for the prevention of chronic disease.
For both the patient and physician, the side effects become the deciding factor. The physician usually chooses the smallest dose in order to avoid toxicity and presumably to achieve some benefit. The patient takes the drug irregularly at best.
As an example, consider the appropriate dosage for statin therapy for the prevention of atherosclerotic cardiovascular disease. Although numerous RCTs have defined the effective dose of a number of statins, recent trends in therapeutics have advocated that rather than using the dose that was used in RCTs, clinicians should increase the dose in order to reach a specific LDL cholesterol blood level.
Choosing the dosage of a drug in an RCT is a less-than-perfect exercise. Here’s how it usually goes:
Phase I trials – often based on pharmacokinetic data derived from animal studies – examine the physiological characteristics of the drug in healthy human volunteers in order to determine an effective and safe dosage prior to a phase II trial.
Phase II trials are larger; they usually examine the effect of several different dosages on a target population, and are focused not on physiological effects but on clinical outcomes and safety, in order to choose the best dosage for a phase III study. Because of their small size, these phase II studies are underpowered and prone to providing misleading dose choices.
Nevertheless, one or two doses are chosen to be used in the definitive phase III RCT, which includes enough patients to provide proof of benefit and safety of the drug based solely on its effect on mortality and morbidity.
Information is often collected in regard to the physiological effects of the drug on, for example, LDL cholesterol (in the case of statins) or heart rate (in the case of beta-blocking drugs). The proof of benefit, however, is determined by clinical outcomes, not on the physiological or "surrogate" measurements.
In the process of designing an RCT, we often make presumptions about mechanisms and will identify certain parameters that theoretically provide insight into the presumed benefit. However, many of the drugs we use have physiological effects that extend beyond the specific therapeutic target. We often remain ignorant about the mechanism by which drugs express their benefit long after their proof of benefit is demonstrated.
Statins, for instance, have a variety of pleiotropic effects. One of the most interesting is their ability to modulate inflammation, a process that is thought to be central to the progression of atherosclerotic disease. Although we presume that their effect is on LDL cholesterol, that presumption may be incorrect. Similarly, beta-blockers have well-known effects on heart rate and blood pressure, but their effect on modulating the up-regulated sympathetic nervous system in heart failure has presumed importance well beyond their effect on heart rate and blood pressure.
It is tempting to make presumptions about the effect of a drug intervention on the basis of surrogate measures like heart rate or LDL cholesterol effects, but their mechanisms of action on mortality and morbidity of disease may be unrelated to that measure.
RCTs have come a long way from relying on "surrogate" end points as the basis for making therapeutic decisions. More than 20 years ago, the CAST (Cardiac Arrhythmia Suppression Trial) was the watershed RCT that excluded the surrogate as a measure of therapeutic efficacy (J. Am. Coll. Cardiol. 1991;18:14-9). At a time when ventricular premature contraction (VPC) suppression was the "mantra" to prevent sudden death, CAST examined the pharmacologic suppression of VPCs in post–MI patients and found that, as the drugs decreased ventricular ectopy, mortality increased.
The use of the seemingly appropriate and obvious "surrogate" of LDL cholesterol lowering as a measure of therapeutic efficacy may be just as illusory. As enticing as surrogates are, the contemporary drive to lower LDL cholesterol may be as misdirected as the target to decrease the frequency of VPCs to prevent sudden death.
Like many things in life and science, things may not be what they seem.
Dr. Goldstein, the medical editor of Cardiology News, is a professor of medicine at Wayne State University and division head emeritus of cardiovascular medicine at Henry Ford Hospital, both in Detroit. He is on data safety monitoring committees for the National Institutes of Health and several pharmaceutical companies.