Diabetes and cardiovascular disease are closely linked, with a high percentage of those with diabetes dying of a cardiovascular illness or event. This is well-recognized, but the relationship between the two diseases is obviously complex.

And the understanding of that relationship got muddier, rather than clearer, with the results of two recent trial looking at aggressive glucose control with the use of drug therapy and its relationship to other factors, such as cardiovascular disease.

Diabetes is associated with a shorter life span, and high blood sugar is a known risk factor for both microvasculature events kidney and eye trouble and macrovasculature events heart attacks and strokes. But it has been unclear how much of an effect lowering blood sugar would have on the risk of heart attack and stroke. Hence, the two trials set out to test whether rigorously bringing blood sugar down to near-normal levels and levels lower than those recommended in current practice guidelines would reduce the number of heart attacks and strokes suffered by diabetics.

In February, the Control Cardiovascular Risk in Diabetes (ACCORD) trial, a large study evaluating the effects of tight glucose control on heart attack and stroke risk was halted due to excess deaths in the treatment arm, which aimed to radically lower blood glucose levels to near-normal levels so-called "tight glucose control."

ACCORD investigators gave a press briefing about their data and conclusions, following that trial halt at the early June annual meeting of the American Diabetes Association (Alexandria, Virginia) in San Francisco in early June, along with investigators from another trial, Action in Diabetes and Vascular Disease (ADVANCE).

The two trials had contradictory results, but the same bottom line: that to reduce the risk of heart attacks and strokes in patients with diabetes, it is more useful to focus on blood pressure (BP) and blood lipid levels than blood sugar. But aside from that, the studies raise as many questions as they answer: the ADVANCE trial did not see the increased risk of death associated with tight blood sugar control; the ACCORD trial produced data indicating increases in cardiovascular disease, that increase leading to the early halting of part of the ACCORD trial.

Further, subsequent analyses have not uncovered any one smoking gun that is responsible for the roughly 20% increase in overall risk of death or the 35% increase in risk of death due to cardiovascular causes in the ACCORD trial.

"At this point in time, none of our analyses have identified any one reason" for the higher death rate, in patients with tight glucose control, Hertzel Gerstein of the ACCORD study group, said at the briefing. Hee said, rather, that it was "the strategy in its totality" that likely was the cause of the higher death rate.

The two trials differ in the details, but both used a mixture of drugs to get patients down to the desired blood glucose level, including thiazoledinediones, sulfonylureas, metformin, and insulin.

ADVANCE showed no sign of the excess deaths that produced a halt in the ACCORD trial's tight control arm cold this winter, but also showed no positive effect of tighter blood glucose control on either cardiovascular complications or overall death rate, though it did show a roughly 20% reduced risk of kidney disease with tight blood sugar control.

The reasons for the differences between the two trials are no clearer than the reasons for the high death rate in ACCORD. At the press briefing, Anushka Patel, ADVANCE director, said that the differences were perhaps not surprising. "These two studies are very different studies, and the populations in the studies are very different," she said.

Investigators noted that participants in the ACCORD trial had poorer blood sugar control to begin with, which also means that they experienced greater changes in blood sugar over the course of the study. Also, more patients in the ACCORD trials used the drug rosaglitazone, which itself has been linked to cardiovascular risk, though there was no difference in rosaglitazone use between the different experimental arms.

Beyond their direct findings, the trials illuminate a larger point about clinical trials: that focusing on risk factors, while often a useful approach and sometimes the only one available, also brings with it the risk of missing the forest for the trees.

Simultaneously with their presentation at the ADA meeting, the trial results were published in the June 6 online edition of the New England Journal of Medicine.

In a Perspectives article that accompanies the papers, Harlan Krumholz and Thomas Lee write that current treatment goals are often formulated in terms of the level of a risk factor without paying any attention to how that level is achieved.

And they point to other recent examples where a trial, despite having the desired effect on a risk factor, showed that a treatment was undesirable in terms of patient outcomes. Hormone replacement in post-menopausal women raises the risk of a heart attack despite lowering bad cholesterol levels; in 2006, phase III trials of Pfizer's cholesterol lowering agent torcetrapib were halted when it became clear that patients receiving the compound had higher death rates despite its desirable effects on levels of both good and bad cholesterol.

"A clinical trial is ultimately a test of a strategy, and we should not be surprised that different strategies may have different effects on patients beyond their effect on risk-factor levels," they write. "ACCORD, ADVANCE and other recent trials remind us that practice is complex and that ultimately we need to understand a strategy's effect on people, not just on surrogate endpoints."

Patel, at the briefing, said that a longer follow-up might still show significant patient benefits as well as surrogate endpoints. Patients have been followed for five years to date in ADVANCE, anfor 3.5 years in d ACCORD before study halt.