To the Editor: In the Action to Control Cardiovascular Riskin Diabetes (ACCORD) study (June 12 issue),1 the intensive-therapygroup differed from the standard-therapy group not only in theglycated hemoglobin level achieved but also in the number ofdrug classes prescribed to achieve it. It is plausible thatin the subgroup of patients with a baseline glycated hemoglobinlevel of 8% or less who had a benefit with intensive treatment,the target glycated hemoglobin level was achieved with fewerdrug classes than in the subgroup of patients with a higherbaseline glycated hemoglobin level. Thus, one wonders whetherthe excess deaths in the intensive-therapy group occurred disproportionatelyamong patients who required more drug classes. Furthermore,was the median glycated hemoglobin level among patients in theintensive-therapy group who died or had cardiovascular eventssimilar to that in the rest of the group? Did the excess deathsin the intensive-therapy group occur in patients in whom thetarget glycated hemoglobin level was achieved or in those inwhom the target was not achieved despite aggressive polypharmacy?Is it possible that instead of achieving a lower level of glycatedhemoglobin, the aggressive polypharmacy that was necessary tocontrol refractory hyperglycemia was associated with excessdeaths?
Elizabeth R. Jenny-Avital, M.D. Jacobi Medical Center Bronx, NY 10461 jennyavita{at}earthlink.net
References
The Action to Control Cardiovascular Risk in Diabetes Study Group. Effects of intensive glucose lowering in type 2 diabetes. N Engl J Med 2008;358:2545-2559. [Free Full Text]
To the Editor: The authors of the report on the ACCORD trialfail to emphasize that the observed increase in mortality withan intensive glucose-lowering strategy could have been relatedto the adverse effects of the drugs chosen rather than due toa lowering of the glycated hemoglobin level to 6.5%. Heavy useof rosiglitazone in the intensive-therapy group raises concern,despite the inability, claimed briefly by the authors, to statisticallyrelate its use to the increased mortality. The fact that fluidretention, weight gain, and congestive heart failure, all potentialadverse effects of rosiglitazone, developed in a higher proportionof patients in the intensive-therapy group supports such a possibility.Two meta-analyses have shown that the use of rosiglitazone wasassociated with an increased risk of myocardial infarction anddeath from cardiovascular causes.1,2 It would be interestingto look at the outcomes separately in study participants takingrosiglitazone in both groups. The public, patients with type2 diabetes, and physicians taking care of such patients needto be correctly informed in order to avoid the premature conclusionthat normalizing glycated hemoglobin to 6.5% is harmful in high-riskpatients with type 2 diabetes.
Fu L. Luan, M.D. Kim Nguyen, M.D. University of Michigan Ann Arbor, MI 48109 fluan{at}med.umich.edu
References
Nissen SE, Wolski K. Effect of rosiglitazone on the risk of myocardial infarction and death from cardiovascular causes. N Engl J Med 2007;356:2457-2471. [Erratum, N Engl J Med 2007;357:100.] [Free Full Text]
Singh S, Loke YK, Furberg CD. Long-term risk of cardiovascular events with rosiglitazone: a meta-analysis. JAMA 2007;298:1189-1195. [Free Full Text]
To the Editor: The ACCORD Study Group draws conclusions thatare surprisingly different from those of the Action in Diabetesand Vascular Disease: Preterax and Diamicron Modified ReleaseControlled Evaluation (ADVANCE) Collaborative Group, reportedin the same issue.1 One obvious difference between these twostudies deserves analysis. The population of patients in theADVANCE trial began and ended the study with a mean weight of77 to 78 kg in all groups. Patients in the ACCORD trial startedout with a mean weight of 93 kg (fully 15 kg heavier), and alarge number of patients in each group gained weight duringthe trial. In fact, 14% of patients in the standard-therapygroup and 28% in the intensive-therapy group gained more than10 kg. This led to a significant difference in the weight outcomesof the treatment groups. Thus, a major risk factor was introduceddisproportionately into the ACCORD trial. It would be of interestto see the ACCORD Study Group do a separate analysis comparingpatients who did not have significant weight gain in each ofthe trial groups to see whether the conclusions hold. Or, conversely,was morbidity or mortality greater among the patients who hadinordinate weight gain in each treatment group?
Ted A. Tobey, M.D. 505 Central Ave. Pacific Grove, CA 93950 ttobey{at}ttobey.com
References
The ADVANCE Collaborative Group. Intensive blood glucose control and vascular outcomes in patients with type 2 diabetes. N Engl J Med 2008;358:2560-2572. [Free Full Text]
To the Editor: In the article by Patel et al. (the ADVANCE CollaborativeGroup), Table 1, which details the baseline characteristicsof the two study populations, does not specify the type of blood-pressuremedications the subjects were taking.
The incidence and progression of nephropathy in subjects withdiabetes have been shown to be related to the type of blood-pressuremedications being used to control hypertension, with angiotensin-converting–enzyme(ACE) inhibitors1,2 and angiotensin-receptor blockers knownto be renoprotective.3 On the other hand, other hypertensionmedications such as amlodipine have not been beneficial in reducingthe progression of microalbuminuria in subjects with diabetes,despite the similar blood-pressure control those drugs provide.4
The type of blood-pressure medication used also affects cardiovascular-relatedmortality and morbidity, especially in subjects with diabeteswho have a history of vascular disease.5 Since the specificclass of hypertensive medications affects nephropathy and otherend points measured in the study, documentation of their usein the two study groups and stratification would have made thestudy results more credible.
Amitabh Parashar, M.D. Carilion Clinic Roanoke, VA 24033 aparashar{at}carilion.com
References
Strippoli GF, Craig MC, Schena FP, Craig JC. Role of blood pressure targets and specific antihypertensive agents used to prevent diabetic nephropathy and delay its progression. J Am Soc Nephrol 2006;17:Suppl 2:S153-S155. [Free Full Text]
Strippoli GF, Craig M, Craig JC. Antihypertensive agents for preventing diabetic kidney disease. Cochrane Database Syst Rev 2005;4:CD004136-CD004136. [Medline]
Coronel F, Cigarrán S, García-Mena M, Herrero JA, Calvo N, Pérez-Flores I. Irbesartan in hypertensive non-diabetic advanced chronic kidney disease: comparative study with ACEI. Nefrologia 2008;28:56-60. [ISI][Medline]
Yasuda G, Ando D, Hirawa N, Umemura S, Tochikubo O. Effects of losartan and amlodipine on urinary albumin excretion and ambulatory blood pressure in hypertensive type 2 diabetic patients with overt nephropathy. Diabetes Care 2005;28:1862-1868. [Free Full Text]
Setoguchi S, Glynn RJ, Avorn J, Mittleman MA, Levin R, Winkelmayer WC. Improvements in long-term mortality after myocardial infarction and increased use of cardiovascular drugs after discharge: a 10-year trend analysis. J Am Coll Cardiol 2008;51:1247-1254. [Free Full Text]
Dr. Byington and colleagues reply: The ACCORD glycemia trialwas designed to determine whether targeting a glycated hemoglobinlevel of less than 6% reduced major cardiovascular outcomesas compared with targeting a glycated hemoglobin level between7 and 7.9% in patients with established type 2 diabetes, othercardiovascular risk factors or cardiovascular disease, and aglycated hemoglobin level of 7.5% or higher. A comprehensivestrategy comprising lifestyle recommendations and all availableclasses of glucose-lowering drugs was used to target the twoglucose levels. Because of the randomized design, the most appropriateconclusions from the ACCORD trial result from the comparisonof these two comprehensive strategies. Whether one or anothercomponent of these strategies was responsible for the findingsis certainly of interest and warrants exploration. However,because participants were randomly assigned to a comprehensivestrategy and not any particular component of the strategy, exploratoryanalyses can only be hypothesis-generating.
The letters by Jenny-Avital, Luan and Nguyen, and Tobey allraise hypotheses that may account for ACCORD's glycemic findings.These and other possible explanations are being explored inmultiple analyses to be submitted for careful peer review andpublication. The common suggestion in these letters is for amore detailed presentation of analyses examining the possiblecause (or causes) of the excess mortality observed in the ACCORDintensive-therapy group. As noted in our article, analyses attemptingto determine the effects of interventional components on outcomes(such as mortality) are confounded by postrandomization characteristicsof the participants or treatment changes. Such analyses mustbe conducted carefully and are challenging to interpret. Nonetheless,the ACCORD Study Group is very interested in these questions,and we are working diligently and systematically to publishour findings in great detail.
Robert P. Byington, Ph.D. Wake Forest University School of Medicine Winston-Salem, NC 27157 bbyingto{at}wfubmc.edu
Hertzel C. Gerstein, M.D., M.Sc. McMaster Medical Centre Hamilton, ON L8N 3Z5, Canada
William T. Friedewald, M.D. Columbia University Mailman School of Public Health New York, NY 10032
Dr. Patel and colleagues reply: Parashar raises concerns thatdifferences in the use of various blood-pressure–loweringdrugs between the intensive-control group and the standard-controlgroup of the ADVANCE trial might explain some of the observedeffects on the study outcomes, particularly with respect tonephropathy. This was not the case. As would be expected ina randomized trial of this size, there were no differences inthe use of such drugs between the randomized groups at baseline.This includes use of ACE inhibitors (43.1% of patients in theintensive-control group and 42.9% in the standard-control group),angiotensin-receptor blockers (5.5% and 5.4%, respectively),beta-blockers (24.3% and 24.7%), thiazide or thiazide-like diuretics(14.4% and 14.2%), and calcium-channel blockers (30.6% and 31.0%).By the end of the follow-up period, the use of such treatmentsremained similar between the intensive-control group and thestandard-control group for ACE inhibitors (52.6% and 50.9%,respectively), angiotensin-receptor blockers (12.4% and 12.6%),beta-blockers (33.7% and 33.6%), thiazide or thiazide-like diuretics(4.3% and 5.1%), and calcium-channel blockers (36.9% and 37.4%).However, we would also contend that there is currently no clearevidence from randomized trials that particular classes of blood-pressure–loweringagents afford different levels of protection against macrovasculardisease in people with diabetes.1
Tobey highlights differences in weight gain between randomizedgroups in the ADVANCE and ACCORD trials and speculates thatthese might account for some of the differences in the findingsof the two studies. Overall, the strategy of intensive glucosecontrol used in the ADVANCE trial was not associated with significantweight gain. However, although it may be reasonable to speculatethat weight change may affect clinical outcomes, adjusted orstratified analyses based on postrandomization weight changeare potentially highly confounded. We do not believe such analyseswould meaningfully contribute to an understanding of the effectsof differences in weight change on the occurrence of the studyoutcomes.
Anushka Patel, M.D., Ph.D. Stephen MacMahon, D.Sc., Ph.D. John Chalmers, M.D., Ph.D. George Institute for International Health Sydney, NSW 2050, Australia apatel{at}george.org.au
References
Blood Pressure Lowering Treatment Trialists' Collaboration. Effects of different blood pressure-lowering regimens on major cardiovascular events in individuals with and without diabetes mellitus: results of prospectively designed overviews of randomized trials. Arch Intern Med 2005;165:1410-1419. [Free Full Text]
Previous
