The excess risk of all-cause mortality associated with intensive treatment in the ACCORD trial was associated with persistently high hemoglobin A_1c rather than lowHbA_1c.

That conclusion comes from a post hoc analysis of data from the Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial, the glycemia arm of which was stopped more than 2 years earlier than planned because of an unexpected 22% excess in mortality in the group randomized to intensive glycemic control aimed at achieving an HbA_1c below 6%.

Several possible mechanisms have been suggested to explain the finding, including hypoglycemia, a rapid lowering of glucose early in the trial, or both. The current analysis suggests that neither of these is correct, Dr. Matthew C. Riddle and his associates reported in the May issue of Diabetes Care.

Overall in ACCORD, all-cause mortality was 1.4% per patient-year in the group of 5,128 randomized to intensive treatment and 1.1% per patient year among the 5,123 randomized to standard therapy. Although the average HbA_1c had plateaued at 6.4% in the intensive group and 7.5% in the standard group after the first year of treatment (from a baseline median of 8.1% in both groups), there was substantial overlap in HbA_1c between the two groups. In fact, average HbA_1c in the intensive treatment group was at least 7% at 40% of study visits.

Deaths from both cardiovascular and noncardiovascular causes occurred over a wide range of HbA_1c values with both treatment strategies, with considerable overlap between the strategies, said Dr. Riddle of Oregon Health and Science University, Portland, and the rest of the ACCORD investigators.

Average HbA_1c was the strongest predictor of death for both groups: A 1-percentage-point increase was associated with a 22% increase in mortality, after adjustment for a variety of potentially confounding baseline factors. But when each group was examined separately, the relationship between HbA_1c and death was much stronger among those in the intensive treatment group, with a statistically significant 66% increase in all-cause mortality for every 1-percentage-point higher HbA_1c, compared with a nonsignificant 14% increase for the standard treatment group (Diabetes Care 2010;33:983-90).

The greatest excess risk of death associated with the intensive treatment group occurred among the patients whose average HbA_1c remained above 7% despite their treatment assignment. Patients who were unable to reduce HbA_1c after initiation of the intensive strategy and continued to have an average HbA_1c above 7% seemed to be at greater risk than did either those with an average HbA_1c of less than 7% with intensive treatment or those with an HbA_1c above 7% using standard therapy.

Factors that would predict a patient’s being unable to achieve an HbA_1c of less than 7% despite intensive treatment remain unknown, the investigators noted.

In two other analyses, the last HbA_1c recorded before death and the decrease in HbA_1c over the first year did not differ between the two groups, suggesting that the rate of change in HbA_1c from baseline was not associated with increased risk of death, they said.

In an accompanying editorial, Dr. Edward J. Boyko commented that these and other recently published data have suggested that three “reasonable assumptions” about the cause of the unexpected ACCORD findings – better glucose control leads to hypoglycemia, hypoglycemia results in excess deaths, and better glucose control is related to higher mortality – are not the case (Diabetes Care 2010;33:1149-50).

“A clue, if there is one to be found, may be the higher mortality associated with resistance to glucose control. ... Whether continuing to pursue intensive treatment in patients in whom [there is] no improvement in glucose control leads to more harm than good should be pursued as a possible explanation for the puzzling findings of ACCORD,” said Dr. Boyko of the U.S. Veterans Affairs Puget Sound Health Care System and the University of Washington, Seattle.

Dr. Riddle has received honoraria from, been a paid consultant for, and/or received research support from Amylin Pharmaceuticals, Eli Lilly and Co., Sanofi-Aventis, Pfizer, and Valeritas. Other members of the ACCORD research team reported similar disclosures. Dr. Boyko is supported by the U.S. Department of Veterans Affairs and the U.S. National Institute of Diabetes and Digestive and Kidney Diseases and did not report any potential conflicts of interest.

Copyright (c) 2009 Elsevier Global Medical News. All rights reserved. This material may not be published, broadcast, rewritten, or redistributed.

控制糖尿病心血管危险行动(ACCORD)试验中强化治疗的过高全因病死风险与持续高(而非低)血红蛋白A1c水平相关。

 

该结论源自对ACCORD数据进行的一项事后分析。该试验的血糖分支研究因强化血糖控制组(目标为使HbA1c降至低于6%)的病死率意外增加22%而提前终止(比计划提前2年以上)。

 

可解释该结果的一些可能机制包括低血糖、试验早期血糖降低迅速或兼有两者。 Matthew C. Riddle博士及其同事在5月份出版的《糖尿病护理》(Diabetes Care)杂志上称，当前分析表明，这两种机制都不对。

 

总体而言，在ACCORD试验中，强化治疗组(n=5,128)和标准治疗组(n=5,123)的全因病死率分别为1.4%/病人年和1.1%/病人年。 基线时，强化治疗组和标准治疗组的中位HbA1c水平均为8.1%。第1年治疗后，两组的平均HbA1c水平分别稳定在6.4%和7.5%。尽管如此，两组的HbA1c水平仍存在明显重叠。事实上，在40%的研究访视中，强化治疗组的平均HbA1c水平至少为7%。

 

波特兰俄勒冈健康与科学大学的Riddle博士及其他ACCORD研究者表示，在两个治疗组明显重叠的广泛HbA1c水平范围内均观察到心血管和非心血管性死亡。

 

在两组中，平均HbA1c水平均为死亡的最强预测因子。对各种基线混杂因素进行校正后发现，每1%的HbA1c水平增幅与22%的病死率增幅相关。但单独分析各组发现，强化治疗组中HbA1c水平与死亡之间的关联更为密切，HbA1c水平每增加1%，则该组病死率显著增加66%，而标准治疗组病死率增幅不显著，仅为14%(Diabetes Care 2010;33:983-90)。

 

在强化治疗组中，平均HbA1c水平在治疗后仍高于7%的患者死亡风险增幅最大。 接受强化治疗后HbA1c水平未见降低且平均HbA1c水平持续高于7%的患者，其死亡风险高于接受强化治疗后平均HbA1c水平低于7%的患者和接受标准治疗后HbA1c水平高于7%的患者。

 

研究者指出，尚不清楚哪些因素可预测患者接受强化治疗后无法达到＜7%的HbA1c水平。

 

他们指出，另外两项分析显示，死亡前记录的末次HbA1c水平和第1年后的HbA1c水平降幅在两组中均无差异，表明HbA1c水平自基线起的变化率与死亡风险增加无关。

 

Edward J. Boyko博士在随刊编者按中评论说，这些数据和其他近期发表的数据表明，以下3种“合理假设”不能用于解释ACCORD试验意外结果的原因：强化血糖控制引起低血糖、低血糖引起过多死亡、(因此)强化血糖控制与较高的病死率相关(Diabetes Care 2010;33:1149-50)。

 

美国退伍军人事务部普吉特湾保健系统和西雅图华盛顿大学的Boyko博士说：“如果存在某种提示的话，那么这种提示就是，较高的病死率可能与血糖治疗抵抗相关……这一令人费解的ACCORD试验结果的出现，是否是因为继续对血糖控制未见改善的患者进行强化治疗弊大于利所致呢？”

 

Riddle博士披露从Amylin Pharmaceuticals、礼来、赛诺菲安万特、辉瑞和Valeritas公司获得酬金、顾问费和(或)研究支持。其他参与ACCORD的研究者报告相似披露。Boyko博士获美国退伍军人事务部和美国国立糖尿病、消化及肾疾病研究所的支持，但未报告任何潜在利益冲突。

 

爱思唯尔  版权所有