PDA

View Full Version : ACCORD update:


xMenace
10-28-2009, 05:03 AM
theheart.org: Cardiology news, educational programming, and opinions (http://www.theheart.org/article/1015749.do)

Lipid/Metabolic
ACCORD update: "One size fits all" may not be best glucose-lowering strategy
October 27, 2009 | Marlene Busko



Montreal, QC - New post hoc analyses of the Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial data suggest that certain baseline factors, as well as the inability to reach treatment targets in a timely manner, may help identify subgroups of patients who are at increased risk from an aggressive glucose-lowering strategy, an ACCORD investigator said at a symposium here at the International Diabetes Federation (IDF) 2009 World Diabetes Congress [1].

"From the glucose-lowering arm of the ACCORD trial, we now know that treating everybody to a very aggressive target probably is not safe, but the analyses that I showed today suggest that there are very likely to be subgroups for whom it is the right treatment, whereas for other subgroups it's not indicated at all," speaker Dr Matthew C Riddle (Oregon Health & Science University, Portland) told heartwire.

Speculating on results from further research, he said that " 'one size fits all'—attacking lipids, blood pressure, and glucose with equal fervor—will probably not be the best treatment approach."




ACCORD details
ACCORD aimed to determine whether intensive glucose control to a target HbA1c of less than 6% would reduce CV outcomes more than a standard target HbA1c of 7% to 7.9%, looking at a primary end point of a composite of CV death, nonfatal MI, or nonfatal stroke.

At baseline, the patients had a mean HbA1c of 8.3% and a relatively long duration of diabetes (about 10 years). They were also moderately obese (average body-mass index of 32), and more than one-third had previous CV events. After one year, patients in both treatment arms attained target HbA1c levels, but at 3.4 years into the planned 5.6 years of the study—as previously reported [2] and widely discussed, the glycemic-lowering arm of ACCORD was stopped due to a 22% excess of all-cause mortality (a secondary outcome) in the intensive arm.

-MB





Riddle presented findings from ongoing post hoc analyses of ACCORD data that tried to tease out which patients might be at greatest risk. Three baseline factors among patients in the intensive-glucose-lowering arm emerged as predictors of increased risk of mortality: HbA1c of 8.5% or greater, neuropathic symptoms, and aspirin use.

"While it's risky to extrapolate what this means, it may be hypothesized that a high HbA1c is a surrogate for relative severity and metabolic control, history of neuropathy is a surrogate for established and significant microvascular disease, and aspirin use may be a surrogate for known or suspected CVD," Riddle speculated.

In addition, patients who were unable to attain target glucose levels within one year also appear to have a higher mortality risk. "If patients start with an HbA1c above 8%, and treatment is intensified, and they are unable to make much progress in lowering it even to the 7%-to-7.5% range, that's a bad sign," Riddle told heartwire.

"These analyses, in our view, do not support the view that rapid reduction of glucose levels or lowering of HbA1c independent of other factors led to an excess risk of all-cause CV death with an intensive strategy," he added. "However, they do not identify an actual cause. We still do not know the mechanisms involved in this unfavorable finding."

Stressing that these findings can be viewed only as hypothesis-generating, he added, "We have a lot of remaining analyses to report, and I hope that the conclusions can be extended."



Not much new, further study of nonresponders needed


Commenting on the presentation to heartwire, session cochair Dr Steven E Kahn (University of Washington, Seattle) stated: "Frankly I don't think there is anything much that was phenomenally new or out of the ordinary" in Riddle's analyses.

"We're still left with the odd finding of the ACCORD study—the increased mortality that has not been confirmed in the Action in Diabetes and Vascular Disease [ADVANCE] study, and interestingly in the recent meta-analysis of the ADVANCE, ACCORD, Veterans Affairs Diabetes Trial [VADT], and United Kingdom Prospective Diabetes Study [UKPDS] data—which all would strongly suggest that glucose control is not associated with major problems in terms of CVD, and clearly people who've had previous CVD are at high risk of having another cardiovascular event."

Like Kahn, Dr Lawrence A Leiter (University of Toronto, ON), an investigator in both ACCORD and ADVANCE, commented to heartwire that the "good news" is that the recent meta-analysis of ACCORD, ADVANCE, VADT, and UKPDS showed that overall, there's no increase in mortality [4].

"The ACCORD study protocol mandated that in every visit, if patients' HbA1C levels were above target, we had to increase the therapy," he said, "and what these analyses would suggest is that this doesn't work and is potentially harmful."

Increased mortality in patients in the intensive-glucose-lowering group occurred among those whose HbA1C did not attain target levels and did not decline quickly. "So it's really not intensive therapy that was associated with the increased mortality, it was unsuccessful intensive therapy," Leiter said.

"To me, the clinical message from all this is 'Yes, we want to get HbA1C down, but at the same time, if you're not succeeding, at a certain point, one has to give up; there's no point in just progressively giving more and more medication if the HbA1C is not coming down.'"

Cautioning that these are hypothesis-generating findings, he added that further study might shed more light on why a subgroup of patients in the intensive-glucose-lowering arm did not respond well and had increased mortality.




ADVANCE subgroup findings differ from ACCORD
In contrast to the findings from ACCORD, further analyses using ADVANCE data suggest that risks and benefits from glucose lowering are uniform across diverse patient subgroups, ADVANCE co-principal investigator Dr John Chalmers (George Institute for International Health and University of Sydney, Australia) said at the same meeting session [3].

Patients in the intensive-glucose-control arm—irrespective of their age, sex, duration of diabetes, and initial HbA1C levels—had similar major reductions in microvascular disease without increased cardiovascular mortality, Chalmers said. "By any index you like, we find a very even spread of benefit and safety," he said.

The clinical message is still that "diabetes is a disease that kills largely through CVD, and it's worth treating the glucose as well as the blood pressure and the lipids," he noted. "The one thing that all the studies agree on is that there's a strong 15% reduction in MI by intensively lowering glucose," as shown in the Diabetologia meta-analysis [4], he noted.

ADVANCE used a "gentle" set of glucose-lowering agents in the intensive glucose-lowering arm—gliclazide (Diamicron, Servier), insulin in less than half the patients, thiazolidinediones in 17% of patients, and metformin, which might explain some of the differences between ADVANCE and ACCORD, Chalmers suggested.

He also drew attention to a promising new risk equation to predict CVD, based on 10 potential risk factors from ADVANCE data. This risk predictor, which remains to be validated, was discussed in another session at this meeting.

"What's interesting in ADVANCE is that using a gliclazide-based therapy achieved similar glycemic targets as in ACCORD, with much less hypoglycemia and less weight gain, and the increase in mortality seen in ACCORD was not observed," Leiter commented to heartwire.

In addition, newer results from ADVANCE provide very good data on the interaction between blood-pressure lowering and glucose lowering, he added. "Not surprisingly," patients who had the greatest benefit were those who attained optimal glucose and blood-pressure measures, Leiter noted.

-MB

fgummett
10-28-2009, 05:49 AM
Interesting read... thanks for posting John.

In summary ACCORD agrees with DF that YMMV :)

---


"The ACCORD study protocol mandated that in every visit, if patients' HbA1C levels were above target, we had to increase the therapy," he said, "and what these analyses would suggest is that this doesn't work and is potentially harmful."

Increased mortality in patients in the intensive-glucose-lowering group occurred among those whose HbA1C did not attain target levels and did not decline quickly. "So it's really not intensive therapy that was associated with the increased mortality, it was unsuccessful intensive therapy," Leiter said.

"To me, the clinical message from all this is 'Yes, we want to get HbA1C down, but at the same time, if you're not succeeding, at a certain point, one has to give up; there's no point in just progressively giving more and more medication if the HbA1C is not coming down.'"I hope the meaning here is that they should try another approach rather than just increasing meds..?

sumi
10-28-2009, 09:08 AM
Thanks for the update, John. It is certainly agonizingly slow and hardly earth shattering. Possibly in another year or two, my Dr. will have read it and no longer be concerned about my A1C.

sarahspins
10-28-2009, 10:12 AM
"To me, the clinical message from all this is 'Yes, we want to get HbA1C down, but at the same time, if you're not succeeding, at a certain point, one has to give up; there's no point in just progressively giving more and more medication if the HbA1C is not coming down.'"

I am not really sure I agree on the message this ultimately has.. I don't like the "give up" wording. I think setting realistic and attainable goals based on individuals is very important.. any doctor and patient who is able to work together to do that, should ultimately be successful in their diabetes management.

REDLAN
10-28-2009, 11:52 AM
a couple of points...

1) the information comes from a presentation at a conference. The information so presented does not experience the same level of examination that peer review is supposed to provide.

2) Beware "post hoc" analysis and especially beware "ongoing post hoc analysis." The problem with post hoc analysis is that the general process works something like this...

Researcher divides the main group up into multiple subgroups, he then tests these subgroups against various parameters, hoping for one of these parameters to score that p<0.05 significance hit. In this case they get hits on Hba1c, neuropathy and aspirin use..... no mention of the subgroups/parameters that weren't significant. The fact is that such associations may have ocurred by chance. which is why the most important thing Riddle states is...

Stressing that these findings can be viewed only as hypothesis-generating

jenb
10-28-2009, 12:25 PM
At the risk of sounding like a total ninny, is anyone really surprised that a "one size fits all" treatment regimen is a flawed strategy? I suggest that Tony get a grant writer, obtain some grant funding and write a paper based on the DF study group. It would be supremely informative and no doubt support the contention that diabetes treatment needs to be highly individualized.

Jen

sumi
10-28-2009, 07:38 PM
Well said, ninny! er , I mean Jen:D

Tribbles
10-28-2009, 08:06 PM
Here is an interesting editorial from the New England Journal of Medicine on this and the ADVANCE trial, Intensive Glycemic Control in the ACCORD and ADVANCE Trials (http://www.medicine.utas.edu.au/teaching/year5/cam515_516/cblpage/cblschedule/diabetes_2/ACCORD-ADVANCE_Editorial_NEJM_2008.pdf). It looks like the ACCORD trial was *very* aggressive and the aim was to get the HbA1c below 6 (the average at the end of the trial was 6.4). The negative outcomes for the intensive arm of the trial is clear but as the editorial states the drug combinations used may well have contributed to this (sulfonylureas AND insulin :confused:), and also the results may be different over a longer timescale.

CaptainMike
10-28-2009, 09:43 PM
Thanks for posting xMenace, I found it somewhat humorously ironic that the presenters' name was "Dr. Riddle" Since that is how I view type 2 a lot of the time.

sanflan
10-28-2009, 10:12 PM
Is it due to the fact that high insulin levels can destroy our bodies? Didn't I read somewhere that when they studied people who lived to be over 100 that they all had lower than average insulin levels? Also longer lives have been correlated to fasting, which in turn could result in lower insulin levels.

I am type two, but I notice that if I take Metformin and don't eat carbs, my bg is pretty low even without insulin. I do wonder if the treatment will end up killing me or destroying my body.


Anyway, just like someone else said Riddle is a good name for this researcher. Sometimes I think I am living a riddle.