Tight Glucose Control Did Not Appear to Increase Mortality Risk
Still No Culprit for Excess Mortality with Tight Glucose Control|
June 10, 2009
SAN FRANCISCO, June 10 -- Lower hemoglobin A1c levels of themselves don't appear to increase mortality risk with intensive glucose control, subanalyses of two major diabetes trials revealed.
Nor was severe hypoglycemia more strongly associated with mortality when treatment was aimed below the standard 7.0% target for hemoglobin A1c, ACCORD and VA Diabetes Trial researchers reported here at the American Diabetes Association meeting.
Ruling out these suspected causes left ACCORD investigators at a loss to explain the 22% excess mortality seen with tight glucose control in their trial.
* Explain to interested patients that intensive glucose control includes therapies targeted to get hemoglobin A1c levels down below 7.0%.
* Note that these studies were presented orally at a conference. These data and conclusions should be considered to be preliminary until published in a peer-reviewed journal.
Both trials along with a third -- ADVANCE -- shook up the diabetes community when negative cardiovascular results for all three emerged at last year's ADA meeting.
The 10,251-patient Action to Control Cardiovascular Risk in Diabetes (ACCORD) study showed no significant reduction in composite nonfatal MI, nonfatal stroke, and death from cardiovascular causes (hazard ratio 0.90, 95% confidence interval 0.78 to 1.04). (See ADA: ACCORD Diabetes Trial a Complete Bust)
Intensive therapy in the 1,791-patient VA Diabetes Trial similarly failed to have a significant impact on cardiovascular event rates (HR 0.868, P=0.12). (See ADA: VA Diabetes Trial Appears to Vindicate Rosiglitazone (Avandia) Safety)
With no apparent cardiovascular benefit in any of the trials and significant harm in ACCORD, tight glucose control was criticized as going too far. (See ADA: Intensive Diabetes Treatment to Blame for Excess Mortality)
However, the latest ACCORD analysis revealed that, paradoxically, lower actually was better in both groups, but particularly for the intensive group, reported Matthew C. Riddle, M.D., of the Oregon Health Science University in Portland.
Each 1% lower average HbA1c was associated with a 56% lower relative risk of mortality after adjustment for other factors in the intensive therapy group (P=0.0001); whereas the same 1% lower HbA1c reduced risk by a nonsignificant 14% in the standard therapy group (P=0.17).
Even more notable was the fact that mortality risk increased steadily for the standard therapy group from a hemoglobin A1c of 6% to 9% but only rose in the intensive therapy group for those with A1c averages above 7%.
"The excess mortality risk in the intensively treated patients occurred with patients who failed to succeed in getting their A1c close to 6% and under 7%," Dr. Riddle said. "It was those who remained high who were at risk."
Thus, it might behoove physicians to back off an intensive goal for patients who cannot achieve a rapid, sustained improvement in A1c with aggressive therapy, which would usually become apparent within three months, he said.
"We don't know why this is," he said. "But we think this has important medical implications" -- that there is a population of patients for whom intensive glucose lowering appears safe.
One important aspect of safety was severe hypoglycemia, which also correlated with mortality, Denise Bonds, M.D., M.P.H., now of the National Heart, Lung and Blood Institute in Bethesda, Md., reported at the packed ADA session.
However, the risk of hypoglycemia was actually lower for the intensive control group patients who got toward goal faster compared with those in the standard therapy group (HR 0.86 versus 0.72% for an initial four-month A1c drop of 2% versus 1%).
The association with mortality was also weaker for hypoglycemia in the intensive group than in the standard group (HR 1.28 versus 2.87). Nor did it account for overall mortality findings, as discussed at ADA last year. (See ADA: Intensive Diabetes Treatment to Blame for Excess Mortality)
Likewise, in the VA Diabetes Trial, severe hypoglycemia was associated with a relatively greater excess risk of cardiovascular mortality and first MI in the standard versus intensive glucose control arm, Stephen N. Davis, M.D., of Vanderbilt University in Nashville, Tenn., reported at the conference.
Although severe hypoglycemia was more common with tight control, there appeared to be no threshold hemoglobin A1c level protective against severe hypoglycemia, which actually tended to increase with elevated A1c, he said.
"There was concern we went too far," commented John Buse, M.D., Ph.D., of the University of North Carolina in Chapel Hill, N.C., and ACCORD steering committee member. "But it's probably more that we pushed too hard in the people who weren't getting to goal."
"You have to individualize these targets," he said, which was the take-home message expressed by many experts and remains part of ADA guidelines.
The VA Diabetes Trial revealed that one important clue to such individualization may be duration of diabetes prior to considering intensive glucose control, reported William C. Duckworth, M.D., of the VA Medical Center in Phoenix.
Diabetes patients had reduced composite cardiovascular event and mortality rates with intensive glucose management if they started on it between four and 15 years after diagnosis.
Risk was elevated, though, with intensive therapy for the rest of patients in a U-shaped curve. Since the VA trial included only patients who'd already failed maximum oral agents or insulin, those who entered the trial in the first three years after diagnosis might be assumed to fit a more difficult to treat profile, Dr. Duckworth said.
Risk of cardiovascular events and death was also somewhat elevated in those who started intensive therapy at least 15 years after diagnosis and was doubled among those who started after 20 years' disease duration.
"The duration of diagnosed diabetes must be a factor in considering initiating intensive glucose control," Dr. Duckworth recommended. "Treat early and treat carefully."
The researchers of both trials cautioned that further research was needed because the studies were not designed to answer these questions and thus the findings they reported here were hypothesis generating only.
ACCORD was funded by the National Institutes of Health, the CDC, and General Clinical Research Centers. Medications, equipment, and supplies were provided by Abbott, Amylin, AstraZeneca, Bayer, Closer Healthcare, GlaxoSmithKline, King, Merck, Novartis, Novo Nordisk, Omron, sanofi-aventis, and Schering-Plough.
The VA Diabetes Trial was supported by the Department of Veterans Affairs.
Dr. Duckworth reported receiving research support from Roche Diagnostics and consulting for Novo Nordisk and Caremark. Dr. Davis reported conflicts of interest with Amylin Pharmaceuticals and sanofi-aventis.
Dr. Riddle reported conflicts of interest with Amylin, Lilly, Sanofi-Aventis, and Pfizer. Dr. Bonds reported no conflicts of interest.
Dr. Buse reported conflicts of interest with Insulet.
Primary source: American Diabetes Association
Riddle MC, et al "New analyses from ACCORD and VADT" ADA 2009.