Early use of low dose Hctz, antidote?
Mere speculation from primary care internist: Clinically suspect increased BPV in DM and have wondered if volume expansion from TZD exacerbates this subtly. This could lead to preferential increased UE FX from protected fall from imbalance. Low dose hctz should help BPV/volume expansion. Fx risk would also be helped with decreased renal Ca excretion. Was Hctz use less in the UE Fx group?

TZD's stink
glad I never jumped on the TZD bandwagon ... now what kind of horror stories will we be seeing with the DPP-4 inhibitors

agree w/ Dr. Ridone
Let's see, TZD's are:
1) the most expensive (until now) oral agents
2) the least efficacious pills (among the "big three" category) at lowering blood sugar

they can also
1) increase the risk of CHF somewhere b/t three and fivefold
2) greatly increase the absolute risk for fractures
3) raise the LDL and increase weight

and most importantly:
No study to date has shown that TZD's decrease overall mortality.

Am I missing something? Maybe a post-hoc subgroup analysis will find a select group of patients that these drugs work well on. :) Keep trying drug companies!

Until then, glyburide, metformin, and insulin remain the standard of care.

Glyburide ? !
SU's would be the last choice if you are trying to reduce the diabetics coronary exit strategy.

not so fast:
From theheart.org's article:

ADOPT: Rosiglitazone delays treatment failure in patients with type 2 diabetics.

(Dr.) Herman said the patients are considered low risk and that the proportion of patients with cardiovascular events was similar in the rosiglitazone and metformin arms but *lower* in the glyburide arm. The rate of congestive heart failure in rosiglitazone- and metformin-treated patients was similar, but *higher* than that associated with glyburide. (The stars are mine.)

So what am I missing? "Lower in the glyburide arm." ". . .higher than that associated with glyburide." If we are trying to prevent CV events, sounds like glyburide is the clear winner here.

Great discussion
David and Steven,
Great discussion. I'd like to add that I believe the answer for the "glitazones" will be in 1. carefully selecting patients who aren't prone to fluid over load---that would be normal EF's, male, those with normal valvular regurg status and normal weight females. 2. primary prevention candidates
3. and for those who can't tolerate ankle edema or are prone to "CHF", the companies need to make available a diuretic combination glitazone formulation.
I'm not certain there is as much "CHF" with these medications as there is ankle edema, and I think that studies are confusing the two, but there is no arguing that these folks can take on fluid like the great titanic. We clearly aren't going where we should be going with these meds in some patients and for others, preventing full blown diabetes and eventualy atherosclerosis is nothing short of a miracle.
Melissa

Physiology and fiction
I agree with Melissa, that we cannot nail the coffin shut until more well selected/designed studies look at the issue.

The mechanism by which TZDs work makes sense that it should help lower CV risk. However, it is still speculation. Between noninferiority trials and primary composite endpoints which include adiponectin levels, it is becoming increasingly cumbersome to sort through the literature these days.

That rant being said, I think in carefully selected patients TZDs do have a role. I still use metformin first +/- glyburide. But with most DM with HTN with creatinine creeping close to 2, it becomes difficult.
My biggest problem with TZD is that fat DM pt who gains 5 pounds, still eats McDonalds and now has an A1c on TZD that is 6.8 says, why do I need to lose weight?

Dan

Adopt trial
Serious CV events with rosiglitazone 3.4 %, glyburide 1.8 % (p< 0.05) ... add this to the fracture problem with rosiglitazone and glyburide does appear as the winner to me also Dave

TZD vs SU ?
Joe and David,you are both on target with ADOPT. My perspective was that of a decade of successful rduction of ACS in diabetics in which TZD's form an integral part of Rx (dovetailed with ACE/ARB, CCB, Hctz,statins, niaspan/tricor,metformin, asa). I do suspect TZD's add benefit in our postprandial world that we do not measure with our fasting parameters. TZD's do facilitate reaching HDL above 45/50 M/F and TG<100. Hypoglycemia has been a very rare complication. I avoid older SU's.

<1/17/6 CMAJ, Simpson analyzed data for 4138 patients with type 2 diabetes taking glyburide and 1537, metformin monotherapy. The authors found that an association between higher daily doses and increased risk of death existed with first-generation sulfonylureas (hazard ratio [HR] 2.1, 95% confidence interval [CI] 1.0– 4.7) and with glyburide (HR 1.3, 95% CI 1.2– 1.4) but not with metformin (HR 0.8, 95% CI 0.7– 1.1). This confirms the findings of previous studies that have also reported an increased risk of death with sulfonylurea drugs. >

This was a preTZD study, but Allen Taylor showed favorable Rosi CIMT effects as did the more recent Pio Chicago Study which showed favorable event trend vs glyburide. Periscope IVUS study with Dr. Nissen may verify this outcome later this year.

thank you for the kind words but,
Periscope IVUS measures a surrogate marker to give us a questionable secondary endpoint - coronary artery atheroma volume. To me, this is yet another example of big pharma inventing a sales pitch.

After the ALLHAT study, Norvasc reps were spouting how well their medicine dilated the carotid arteries in comparison to lisinopril. "So what," I said, "do you have hard endpoints? Does Norvasc prevent more strokes than other drugs?"
Of course, they had no answer.

There are only four things that matter when it comes to medicines: Do they reduce mortality, morbidity, or increase the quality of life?

And if the above three are relatively equal, for societal and patient benefit (which is why we are in our professions) we need to choose the cheapest option.

Drug companies and researchers have a moral obligation to do direct comparisons with competitor drugs involving the three hard endpoints -- morbidity, mortality, and quality of life. Everything else is just window dressing.

To me, the ADOPT study involved direct comparisons, and as far as CV events, glyburide came out the winner. This was not a meta analysis. (I never "meta analysis" I liked.) At the very least, glyburide was non-inferior.

I appreciate being involved in the debate.
And I'm glad we have this forum. Agree or disagree, I always learn something. Thanks all.

ADOPT not a cardiovascular endpoint study
ADOPT was not powered for cardiovascular events -- therefore, we must be cautious when looking at its data on CV event differences. In particular, because it was a comparison of 3 different agents, it would have needed a much greater sample size to declare equivalence between any two comparators on cardiovascular events. Such a trial would have needed tens of thousands of patients as in ALLHAT (which was over 40k with the doxazosin arm). Point seems to be lost in this discussion of ADOPT.

nevertheless
Nevertheless, the CV endpoints were statistically significant (at least the TZD vs. glyburide) and were reported.

And if the study wasn't powered to look at CV endpoints, then we might as well throw out the unexpected data on fractures.

I agree though, we need a study involving direct comparisons with large populations. An ALLHAT-type study involving the "big three" oral DM meds would be great.

I disagree
Underpowered trials often give spurious results. We tend to think in terms of false negatives (not finding a true association or relationship where one is present, because of insufficient sample size or event rates), but the converse are also increased -- false positives, that is, finding a relationship that is not really there.

By shifting the focus from the primary outcome to secondary (CV) and tertiary (subtype of adverse event -- fracture), we are undermining the basis for the trial, which was to look at glycemia. I view the other outcomes as interesting, possibly hypothesis-generating, but woefully underpowered. They need to be the basis of their own primary event driven trials.

Until that time, it would be better to rely on PROACTIVE and UKPDS in making points about these other merits (or lack thereof) of antiglycemic agents.

The other problem -- too many types of outcomes and focusing only on a few (eg fracture) -- raises the possibility of multiple hypothesis testing and coming up with findings by chance alone.

Perceptions
Interesting commentary here.

I think it is important to keep in mind what the primary and secondary endpoints are and use observational endpoints as idea generating (even if statistical). Cases in point:

1. ASPEN lipitor 10 mg showed no statistical benefit last year in people with DM. hmmm should we stop lipitor in people with DM? Lipitor has never shown mortality benefit in studes, is this unique to lipitor or better medical mgmt. we don't know.

2. PROactive (pioglit) 3 year trial showed reduction in mortality, stroke and nonfatal MI but did not reach the primary all vascular event reduction including pvd and amputations in people with DM (for 10 years) over age 60. It did prevent the need for using insulin nearly 60% with an NNT of 9, NNT to prevent the nonstatistical primary endpoint of 42, NNT of 48 for secondary endpoint above and NNT of ~ 60 for sx of CHF or hospital, but no CHF death. Statistically more men had renal cancer on Pio but statistically LESS women had breast cancer on PIO. Does Pio cause renal ca in men and prevent breast ca in women? That is quite a speculation when less than 20 people had these events in a study with several thousand people.

3. UKPDS showed mortality reduction in those pts initiated on metformin, but those who started on SU and had metformin added 6-7 years later the mortality rate was 86% higher. hmmm, interesting, does adding metformin to a baseline of SU cause higher mortality? Remember failure in that study was glucose control and was defined as fasting glucose of 270 mg/dl.

4. ADOPT was a study measuring monotherapy time to failure either fasting glucose sustained above 140 mg/dl or 180 mg/dl. SU showed maximal benefit at 26 weeks and then trend to failure reached at 33 months >35%, metformin at 45 months >20% and TZD at 60 months > 15% (NNT vs. SU 6) . CV events and other events were captured but the Study endpoint was failure for monotherapy control and remember once you failed you were out of the study and no longer monitored. There were several hundred people more in the SU group that were no longer in the study for CV event monitoring, hmmmm.... Less people being monitored for events, less events.

5. CHICAGO showed less CV events in the TZD group than then SU group over 1.5 years. THis was observational as well. Multiple studies now have shown TZD/Glinides reducing athero progression and inducing regression, where to date no SU, Met, Insulin has shown this, perhaps they take longer and we don't know if this translates into mort, morb or QOL benefits.

THE ADA/EASD rec starting with TLC's and Metformin therapy upon dx of DM2 and in 3 months if A1c goal is not reached the second step should be: Insulin as it is the most rapidly effective but can cause hypo and can be complicated to use OR SU inexpensive but hypos OR TZD expensive but low risk hypo but fluid retention monitoring. At 6 months add third agent.

The bone issue with TZD's is interesting. I think the high rates of metabolic syndrome, perimenopause, osteopenia, low ca/mg consumption. high rates of Vit D deficiency, high sodium, low potassium consumption, high diuretic use, etc... are all counfounding issues that likely play a role.

No one can discount the unique issues that metformin, su, insulin or TZD's offer mgmt for glucose and diabetes control in this high risk population. And if a person can't prevent CHF or identify it easily then a TZD may not be appropriate. If one can't monitor renal function then Metformin..... SU's and Insulin if one doesn't understand fasting and postprandial glucose control and instruct pts on carb counting and glucose monitoring. Choose your treatment modality.

In our office we are fortunate to have all of these options available as well as newer agents for those patients that have failed the above modalities. Good Luck.