As physicians, we often rely on surrogate markers to determine the best course of treatment for a given patient. Indicators such as blood glucose, low-density lipoprotein (LDL)-cholesterol and C-reactive protein guide us in day-to-day practice. Some markers have been helpful; others have been misleading. They're often used as surrogate clinical endpoints in trials, generating a flood of data that has left many clinicians confused. While it's clear that lowering LDL-cholesterol and reducing blood pressure saves lives, there are other markers where benefit is less evident.

For instance, premature ventricular contraction (PVC) is a marker of sudden death and ventricular arrhythmias. PVC suppression with antiarrhythmic agents, however, can be deadly; the drugs are only used to improve the quality of patients' lives. Another good example of a surrogate marker gone wrong followed the recognition that hormone replacement therapy (HRT) lowers LDL-cholesterol. Trials later found that HRT increases cardiovascular (CV) events in postmenopausal women and raises the risk of certain cancers such as breast tumours. It also boosts the risk of deep vein thrombosis and pulmonary embolisms.

Clamping down on blood sugar
Over the last year, two large clinical trials looking at blood sugar management have drawn our at-tention. As blood glucose goes up, CV events, neuropathy, eye complications and renal complications all increase, so aggressively lowering blood sugar seems like a reasonable thing to do. Yet the ACCORD (Action to Control Cardiovascular Risk in Diabetes) study, which involved 10,251 diabetic patients, found that intense lowering of glycosylated hemoglobin (HbA1c) to 6.4% - which is below current recommendations - was associated with increased mortality. The treatment arm was stopped in February 2008 after it became apparent that the relative risk of death was up by 22%.
The ADVANCE trial in Europe, however, involving > 11,000 patients with type 2 diabetes and us-ing different blood sugar-lowering agents, found different results. The researchers recently reported that bringing mean Hb_A1c levels down to 6.5% was safe and reduced the risk of CV death, though this observation wasn't statistically significant. They also detected a significant 10% decrease in the over-all risk of serious diabetes complications in this treatment arm.

Balancing the risks
Let's keep in mind that tight blood sugar control is difficult to achieve and leads to side effects. These two large randomized controlled trials haven't been able to show that aggressive glucose lower-ing for periods of up to 5 years provides vascular protection for middle-aged individuals with type 2 diabetes. If anything, the ACCORD trial suggests the opposite: that aggressive blood sugar manage-ment could cause harm. Microvascular complications may respond differently than the standard CV indicators but we need more data on the details in order to weigh the risks and benefits.

Next month: Spotlight on HDL-cholesterol