Home / Resources / Articles / Scores Help Predict Benefit from Intensive Diabetes Therapy

Scores Help Predict Benefit from Intensive Diabetes Therapy

Aug 18, 2011

Validated measures of cardiovascular risk and comorbidity at baseline helped predict response to intensive glucose control in a post hoc analysis of patients who participated in the Veterans Affairs Diabetes Trial….

In the overall VADT, in which 1,791 military veterans with sub-optimally controlled type 2 diabetes were randomized to receive either intensive or standard glucose control, intensive therapy had no significant impact on the rates of major cardiovascular events at a median follow-up of 5.6 years (N. Engl. J. Med. 2009;360:129-39). However, a subgroup analysis of 301 trial participants showed that intensive glucose lowering did significantly reduce cardiovascular events among those with less-extensive calcified coronary atherosclerosis (Diabetes 2009;58:2642-8).

Dr. Nalurporn Chokrungvaranon, an endocrine fellow with the Phoenix VA Health Care System, stated that, “Because coronary artery calcification measurements are expensive, involve radiation exposure, and are not readily available everywhere, it would be advantageous to have alternative clinical indices for predicting which patients would be more likely to benefit from intensive glycemic therapy.”

In the current post hoc analysis, patients were divided into upper, middle, and lower tertiles of scores on four different validated measures: the Framingham 10-year cardiovascular disease (CVD) Risk Score, the U.K. Prospective Diabetes Study score, the Charlson comorbidity index, and another measure for predicting 4-year mortality based on age, sex, self-reported comorbid conditions, and functional measures.  

At baseline, the study patients had a mean age of 60 years and a mean hemoglobin A1c level of 9.4%. Those in the upper tertiles for any of the four scales at baseline showed no benefit from intensive glycemic therapy, with hazard ratios ranging from 0.92 for the 4-year mortality score to 1.06 for the Framingham score.

However, there were differences at the lower and middle tertiles. For the Framingham score, patients in the middle tertile had significantly reduced risk of cardiovascular events with intensive therapy (hazard ratio, 0.66; P less than .05). She noted, however, that the lower tertile did not show that benefit (HR, 0.85).

Similarly, the middle tertile on the Charlson index also showed a significantly lower risk for CVD events (HR, 0.57; P less than .05), whereas the lowest tertile did not (HR, 0.99).

It’s not entirely clear why the lowest tertiles did not show benefit, but event rates were lower in that group, so it could be a power issue. Moreover, CVD events take longer to occur in those at lower risk, so it’s possible that differences would emerge if the study were carried out on a longer timeline, Dr. Chokrungvaranon commented.

Indeed, for the 4-year mortality prediction score, the lowest tertile did show a significantly lower event rate (HR, 0.71; P less than .05). Adjustment for prior CVD did not change the results, she said.

“Cardiovascular scores and comorbidity indices may be useful tools to identify patients who should be considered for less aggressive treatment for diabetes…. These results further support the notion that A1c goals should be individualized and should not be one size fits all,” she concluded.

JAMA 2006;295:801-8