Podcast: Play in new window | Download (12.5MB)
In episode 62, the eighth and final podcast in our tome on the treatment of type-2 diabetes, we talk about cholesterol and ASA. Statins have evidence of benefit, around 3-4% absolute risk reductions over 5 years, other drugs for cholesterol have little if any evidence of benefit, and there is evidence of no benefit from taking ASA. We come up with an overall synopsis; increase activity, eat good food, use metformin, then maybe sulfonylureas, thiazides/ACE inhibitors for BP, statins but don’t measure cholesterol, and no ASA. Both James and Mike collapse from exhaustion and leave the final word to Chris their producer.
Show Notes
1) Statin meta-analyses
Lancet 2008;371:117-25
2) Fibric acid derivatives
FIELD
Lancet 2005;366:1849-61
Fibrate meta-analysis
Arch Intern Med 2005;165:725-30
3) ASA – no benefit in diabetics
{ 11 comments… read them below or add one }
In your brilliant article on BMJ http://www.bmj.com/cgi/content/extract/320/7251/1720 , you said in table 1 that ARR with intensive sugar control for 10 yrs would be 3.2% in comparison with diet advice only. This makes NNT=31 for 10 yrs. Seems like you divided the total number of primary end point by total number of each group; for example, 35.3%= 963/2729 for intensive group (dietary advice plus chlorpropamide, glibenclamide, insulin) and 38.5%= 438/1138 for conventional group (diet only)
Any diabetes related end points* (%)
Dietary advice plus chlorpropamide, glibenclamide, or insulin 35.3
Dietary advice only 38.5
Relative risk reduction 8.2
Absolute risk reduction 3.2
No needed to treat for 10 years to prevent one event 31
However, another article on BMJ http://www.bmj.com/cgi/content/abstract/327/7409/266
regarding re-analysis of UKPDS trials, the authors simply used the UKPDS data and said in table 2 that intensive group 40.9 events per 1000 patient years vs diet only group 46.0 events per 1000 patient years, which makes ARR=0.51 per year and NNT per year=196.
I can’t figure out this huge difference of ARR in these two analyses. Both of them put down the number for 10 year treatment period. I’d appreciate if you can show me some guide on my confusion.
Hi nevermind – thanks for the compliment on the article – all we did was look at the numbers and published what was found. I’m not sure how they got their numbers – you are correct in how we got our numbers. They end up with an ARR of 0.51 per year and we get an ARR of 0.32/year – 3.2/10 years. The difference likely comes from the fact we did a very simple calculation of dividing total number of endpoints by the total number and as every patient isn’t in the study for exactly 10 years they used a per 1000 patient years for their numbers, but I don’t know. Not sure if a 5% versus a 3% absolute reduction over 10 years is all that clinically different in the grand scheme of things – especially when it was a composite of 21 endpoints.
Great podcast, carry on!
A few observations/pet peeves.
1. Thin older patients with type 2 DM who have symptoms related to blood sugar will likely require insulin. Don’t mess around too long with metformin!
2. Completely agree with comments on rosiglitazone/pioglitazone. What a waste of good lab rats that was! Really, now!
3. The (irritating) practice of giving short acting insulin by sliding scale in otherwise stable hospital-bound patients is an overt invitation to hypoglycemia. Better to switch the patient over to an intermediate acting preparation fairly quickly, if one must continue with insulin treatment.
4. Agree with comments on insulin preparations – little to choose from them, stick with NPH, with one exception – our dialysis patients. They seem to do very well on insulin glargine (“Lantus”), especially the ones on peritoneal. Remember, the patient will absorb about 50 gm of glucose with each dialysis, if glucose if used as the osmotic agent in the dialysate. I will not discuss the finer points of which glucometers one should use as this is getting bit too technical – but of real concern for our dialysis Nursing staff. The end points in these patients are: the patient has no symptoms related to elevated glucose, there are no episodes of hypoglycemia, and the average CBG is less than 11 (yes, that is correct, 11 mMol/l, or 200 mg/dL for old hounds like me).
5. A quick reminder of basic physiology. Insulin is secreted into the portal venous system,and ends up in hepatic sinusoids, where 90% (indeed,90%) of the insulin is metabolized by insulinases. Ergo, the major physiologic target of insulin is the liver. Glycogen synthesis,and all that.
Insulin administered subcutaneously or intravenously will primarily hit muscle and fat tissue. It does lower blood sugar, typically by increasing fat stores, and muscle glycogen. This is not repeat NOT physiologically the same to endogenous insulin.
The density of insulin receptors on adipocytes decreases as cellular lipid stores increase. Therefore, the more the weight gain, the less adipose tissue is sensitive to insulin, and the more difficult blood sugar control becomes. This is one of the mechanisms underlying so-called insulin resistance.
Regards,
Jean Chouinard MD
Medical Director
St Vincent Hospital
Ottawa, ON
Dear TEC podcast
Somewhere contained within the vast collection of diabetes podcasts – perhaps this very episode 62 – James made reference to homocysteine-lowering and CVD as yet another good example whereby successful surrogate manipulation fails to impart clinical benefit.
Perhaps this comment deserves differentiation between the cerebral and non-cerebral components of CVD. The HOPE-2 group reported a 25% RR reduction of incident stroke (hazard ratio [HR], 0.75; 95% CI, 0.59–0.97) in the group provided homocysteine-lowering therapy (B6, B12, folate). In a subsequent article, this time specific to stroke, they reported a reduction in nonfatal stroke (HR, 0.72; 95% CI, 0.54–0.95). I calculated a NNT of 78 for prevention of incident stroke.
These findings are of course subject to bias (sub-group and perhaps post hoc analyses), and the modest benefit is insufficient to support routine B vitamin use as a public health strategy. However, I think it’s premature to imply that homocysteine-lowering is without some CVD benefit. Moreover, the discussion deserves that the components of CVD be differentiated.
Thanks for these top-notch podcasts – both of you shining examples of “The Good Doctor”.
Hal Huff BA, ND, MSc (cand)
Associate Professor, Clinic Supervisor
Chief Supervisor Sherbourne Health Centre Community PHA Naturopathic Clinic
Canadian College of Naturopathic Medicine
Toronto, Canada
Gustavo Saposnik, MD, MSc, FAHA; Joel G. Ray, MD, MSc; Patrick Sheridan, MSc; Matthew McQueen, MD, PhD; Eva Lonn, MD, MSc the HOPE 2 Investigators: Homocysteine-Lowering Therapy and Stroke Risk, Severity, and Disability: Additional Findings From the HOPE 2 Trial, Stroke. 2009;40:1365-1372
Hi Hal: Thanks for the kind comments about our podcast and also for your thoughtful comments. Here are some thoughts
A) Issues with HOPE-2 and stroke outcomes
1) stroke on it’s own wasn’t a specified endpoint
2) the HOPE-2 author’s comment on the stroke issue “The number of strokes in our study was much lower than the number of coronary events, the confidence intervals around the estimated risk reduction are wide, and the results are not adjusted for the multiplicity of outcomes compared. Also, we found no effect of treatment on transient ischemic attacks. From a biologic perspective, a treatment benefit restricted to stroke would be difficult to explain.”
NEJM 2006;354:1567-77
3) We hadn’t seen the stroke article you mentioned – thanks – as you state it was a retrospective subset-analysis – and, the absolute reduction was 1.3%. Interestingly despite this difference, the author’s state there was no difference in severity or disability.
We heard one stats guy say, if they don’t correct for multiple analysis (which they didn’t), multiply the p-value by 10. If it is still significant, it may well be real.
Stroke 2009;40:1365-72
B) Other studies
1) Treatment with B vitamins did not lower the risk of recurrent cardiovascular disease after acute myocardial infarction – no effect on stroke and a suggestion of a worse overall outcome
NEJM 2006;354:1578-88
2) No change in stroke rate between subjects given high versus low doses after a stroke
JAMA 2004;291:565-75
3) Meta-analysis of folic acid studies including HOPE-2 found no reduction in stroke
JAMA 2006;296:2720-6
4) One year post stent for 550 people: the composite end point was significantly lower at 1 year in patients treated with homocysteine-lowering therapy (15.4% vs 22.8%; relative risk [RR], 0.68; 95% confidence interval [CI], 0.48-0.96; P =.03)
JAMA 2002;288:973-9
5) After stents, worse outcomes with folate (630 pts x 6 months). “The restenosis rate was higher in the folate group than in the placebo group (34.5 percent vs. 26.5 percent, P=0.05), and a higher percentage of patients in the folate group required repeated target-vessel revascularization (15.8 percent vs. 10.6percent, P=0.05).”
N Engl J Med 2004;350:2673-81
6) No help with stable coronary disease.
J Am Coll Cardiol. 2003; 41: 2105-13
C) The evidence is a dogs breakfast. It is not looking clear either way and the jury is still out. Looking at one outcome in isolation is tricky as one decides to take something usually based on the overall benefit. There is a possibility (in other words we can’t rule out with the evidence we have to date) that homocysteine-lowering (more accurately taking folic acid and vitamin B) may produce a roughly 10-15% relative reduction in strokes. There is a Cochrane protocol for the effects of these agents on CVD – let’s wait and see what that data shows.
THANKS AGAIN
Thanks Jean – great points and very much agree – thanks also for the positive comments about our podcast.
Hi Guys,
I love your podcasts. However, I’m a little confused (as you are) about what our goal A1C’s are with diabetics. You seem to mention that tight control might be good for “new” diabetics. What does this mean practically? If I have a new diabetic, do I put them on metformin and insulin to get their A1c to 6? And when does a diabetic become “old”?
You also seem to mention that you add another drug when the A1c starts creeping up over 8. But, the VA study found no difference in mortality between the intensive arm at 6.9% and the standard arm at 8.4%. Why not just keep everyone under 9%?
I guess we will really never know the optimal target and will have to use risk tools to communicate with our patients. Just like cholesterol and blood pressure.
Hi Rocky3200 – Sorry for the delay in our response – the following is Mike’s comment.
First, about the evidence: the UKPDS was in early DM and the other studies were in established DM. Therefore, if you are going to treat for benefit, better do it more early (UKPDS found more benefit than any of the others). In general, in keeping with UKPDS, I am a little tighter at the beginning of DM. I give no sugar drug until the HgbA1C creeps up around 7.0 then I add metformin (250 BID then up to 500 BID). I then increase Metformin to as HgbA1c is in the low 7.0 to keep it in that range (usually creeping up over a few yrs). Then, I might let things slide a little to the upper 7 or at 8.0, then I consider a second agent. As Mike K said, there is a creep over time and I am not rigid about HgbA1C. The other risks (BP and maybe Chol and certainly pushing all lifestyle) are being addressed. Once they have had DM for a while and the HgbA1C is higher (let say low 8’s), I am just trying to keep them in the 7’s (that means 7-8) but as it really progresses, low 8’s are also good. Balancing the amount of drugs vs the amount of HgbA1C is a real uncertainty.
One of the issues here is talking about these HgbA1C numbers. When we say, “in the 7’s” we don’t mean 7.0, we mean 7-8 (or more correctly, 7.9). When we say something like below 8, it doesn’t mean 7.0, it means 7.8 or something like that. When I say I take action (give a drug) at around 7 to keep just below there, I am talking about 6.8 or 6.9 etc. Now, my attempts don’t match my achievement (a sad truth larger than this question): therefore, the patient might go to 6.5 or just stay at 7.0. If they are 6.5 I probably won’t need do anything for a couple of years but if it is 7, I might need to modify meds in a year. All of that, for the brand new DM patient. Most of mine have been sitting DM for a while and they are in around 8.0 but some who are around a real long time are 8.5 or a little higher (for non-adherent, I think the highest in our clinic was 13). The problem here is people want exact numbers (Precision in a imprecise world). Even if my goal was 7.8, I could not get most of my patients there: some would be 7.7 and some 7.9, etc. The guidelines give exact numbers. Studies and life do not. God, I hope we cure Diabetes.
James states that there is a 6 – 7 % benefit over 5 years in those taking statins with CVD. However, in our course notes (I am a naturopathic doctor) and during the lecture, the value is 4%. Which is correct? Also, where I can I find the percentage of risk for CVD created from: HBP, diabetes, gender, age, etc.?
Thanks.
Hi Graham. It depends what data you look at – I think I presented from the table at http://therapeuticseducation.org/useful-tools/statin-overview/ in the course – in that table there is a 4% reduction in CHD, a 1% reduction in stroke and a 2% reduction in mortality so I should have said the benefit is 5-6%. When we looked at the data at the TI a number of years ago the numbers were between a 6 and 8% absolute benefit see http://www.ti.ubc.ca/node/76. It really depends on exactly what end points are included. The risk for CVD etc in BP and other hopefully useful tables can be found at http://therapeuticseducation.org/useful-tools/. Thanks.
Thank you. That was the exact data I was looking for.
I really enjoy listening to these podcasts and attending the lectures. Before listening and attending, I thought I was in for very boring material I didn’t care to learn. Surprisingly, the podcasts and lectures have been anything by boring, very informative, and a pleasure to listen to and learn from. Great job!
Thanks.
You must log in to post a comment.