3rd August 2009
A meta-analysis of prospective randomised controlled trials suggests a small benefit of intensive glucose control in people with type 2 diabetes in reducing the risk of cardiovascular disease (but not stroke, death from cardiovascular disease or death from all causes). However, intensive glucose control doubles the risk of severe hypoglycaemia. As we have concluded previously, it remains uncertain whether intensive glucose control (e.g. the addition of hypoglycaemic drugs to reduce HbA1c to levels significantly below that of standard treatments) offers any significant benefit in addition to that achievable by successful implementation of other interventions to reduce cardiovascular risk (i.e. smoking cessation, exercise, losing weight, controlling blood pressure, lowering cholesterol, taking metformin).
Level of evidence
Level 2 (limited-quality patient-oriented evidence according to the SORT criteria.
This analysis confirms that, for people with type 2 diabetes, priority should still be given to lifestyle interventions (stopping smoking, losing weight, and taking more exercise as appropriate), controlling blood pressure, taking a statin, and taking metformin. In some patients, additional hypoglycaemic drugs may be considered to control blood glucose. However, NICE guidance should be followed and individual targets for HbA1c should be agreed with each patient. These could be above that of 6.5% (48mmol/mol) set for people with type 2 diabetes in general and should take into account patient preferences and the balance of likely benefits and harms (such as hypoglycaemia) as well as the medicines management issues.
What is the background to this?
There has been considerable debate in recent years with regard to the importance of intensive glycaemic control relative to other measures for reducing cardiovascular (CV) risk in people with type 2 diabetes. A recent MeReC Rapid Review Blog discussed a meta-analysis (MA) of randomised controlled trials (RCTs) that suggested a small benefit of intensive glucose-lowering therapy in reducing coronary heart disease (but not stroke or death) compared with standard therapy. However, the benefit did not appear to be as great as that achieved by blood pressure control or lipid lowering. This MA of five prospective RCTs in 27,802 adults with type 2 diabetes compared the effects of intensive glucose-lowering therapy with standard therapy, and included clinical CV disease as the primary outcome. This MA did not include the PRO-ACTIVE study, which was included in the previous MA. It did include the two UKPDS studies (looking at glycaemic control with sulphonylureas or insulin, or with metformin) separately, whereas they had been pooled before entry into the analysis in the previous MA
What does this study claim?
The authors found that intensive glucose control reduced the risk of CV disease (relative risk [RR] 0.90, 95% confidence interval [CI] 0.83 to 0.98) and coronary heart disease (RR 0.89, 95%CI 0.81 to 0.96). However, more specifically, there was a reduction in nonfatal myocardial infarction (MI) (RR 0.84, 95%CI 0.75 to 0.94) but not in fatal MI, nonfatal stroke, fatal stroke, congestive heart failure or peripheral arterial disease. Intensive glucose control did not reduce the risk of CV death (RR 0.97, 95%CI 0.76 to 1.24) or all-cause mortality (RR 0.98, 95%CI 0.84 to 1.15), However, the risk of severe hypoglycaemia was doubled with intensive treatment (RR 2.03 95%CI 1.46 to 2.81). Over five years, for every 1,000 patients on intensive glucose control, the authors estimated this equates to 15 fewer CV events compared with standard therapy, although this is offset by 39 excess episodes of severe hypoglycaemia.
How does this study relate to other studies?
A previous blog points out the conflicting nature of evidence from RCTs with regard to the benefits and risks of intensive glucose control. The ACCORD, ADVANCE and VADT RCTs, all of which were included in the present MA, have not identified consistently a significant benefit of intensive glycaemic control in the treatment of type 2 diabetes (see also blogs 258, 147 and 64 for more details) with regard to CV outcomes and mortality. Indeed, the ACCORD study was stopped early because of an increased risk of death in the intensive treatment arm.
We have consistently pointed out in other blogs and NPC materials on type 2 diabetes the need to consider glycaemic control in the context of other important interventions (both lifestyle and drug interventions) to reduce CV risk. Potential benefits on macrovascular and microvascular outcomes, and the risks from treatment, need to be considered. This MA suggests that intensive glycaemic control therapy may have some beneficial effect on the risk of CV disease, particularly nonfatal MI, although it would appear to have no significant effect on reducing strokes or all-cause mortality. This needs to be balanced against the increased risk of severe hypoglycaemia associated with intensive glycaemic control.
In addition, there are several aspects of this MA which limit the generalisation of these findings to clinical practice. A meta-analysis of RCTs is often considered to be one of the highest forms of evidence. However, as discussed in our blog of a previous, similar MA, the trials combined are very different, i.e. there was considerable heterogeneity among the studies. There were marked differences in the patient baseline characteristics (e.g. presence of CV disease, duration of diabetes), the treatment protocols (e.g. drugs and doses used, HbA1c targets) and duration of follow-up. In particular, three of the studies included (ADVANCE, ACCORD and VADT) had a roughly similar design in that they examined outcomes after reducing HbA1c to below 7% for several years in typical cohorts of type 2 patients aged around 65+ who had long-standing diabetes. The other two, UKPDS studies, are very different and were in people with newly diagnosed diabetes. For the two outcomes where intensive glycaemic control was found to be beneficial, the only trial which individually found a statistically significant benefit was UKPDS 34, which tested metformin against conventional control, mainly diet, in newly diagnosed people and in which the post-intervention median HbA1c was similar to or higher than that in the control arm in ACCORD, ADVANCE and VADT.
The authors suggest possible heterogeneity between the results of the subgroups analyses of the different trials for the effect of intensive glucose control on CV mortality and all-cause mortality. When the results of the early UKPDS trials were pooled, non-statistically significant protective effects were seen on these outcomes. However, the summary data from the more recent trials (ADVANCE, ACCORD and VADT) showed non-statistically significant increases in the risks of both CV mortality and all-cause mortality with intensive glucose control. These trials used more stringent intensive glucose control and were also looking at a different patient population. The authors suggested that these differences might be partly due to the excess mortality that was found in the ACCORD study. They also noted that the more recent trials included treatment with glizazones, which can increase the risk of heart failure, double the risk of bone fracture in women and there is some evidence to suggest that rosiglitazone, in particular, may be associated with an increased risk of myocardial infarction. In addition, it has been suggested that the excess mortality in ACCORD may have been due to deaths from severe hypoglycaemia. However, it is possible that some deaths in recent trials may have been incorrectly labelled as being ‘due to hypoglycaemia’.
Other limitations of this MA include the limited power of the subgroup analyses to detect small effects of intensive glucose control and the use of summary data from trials rather than individual patient data. Also, the recent trials may have been too short to show effects on CV and total mortality.
The benefits of glycaemic control on coronary outcomes appear modest in comparison with the benefits that might be obtained by smoking cessation and other lifestyle measures, blood pressure control and lipid modification. We have previously pointed out that the benefits to be obtained by intensive glycaemic control are less than those achieved from a 4mmHg reduction in blood pressure or a 1mmol/L reduction in LDL-cholesterol. This emphasises the need to consider a general approach to reduce CV risk in people with type 2 diabetes, and not to focus solely on reducing blood glucose levels.
It is important to consider the message from the three most recent trials (ADVANCE, ACCORD and VADT) included in this MA, which suggests, that lowering HbA1c below 7 in this large group has no clear benefit and increases hypoglycaemia. Intensive glycaemic control can be considered, where necessary, in addition to other interventions for reducing CV risk, but any potential benefits need to be considered against the increased risk of hypoglycaemia and drug-specific adverse effects
More information on the management of type 2 diabetes can be found on the relevant section of NPC
Design: Meta-analysis of prospective RCTs where the sample size was 500 patients or more. Patients: 27,802 patients aged at least 19 years with type 2 diabetes; mean time since diagnosis of diabetes and median HbA1c varied widely between studies; mean BMI 28 to 32 kg/m2. Studies were retrieved systematically by searching the MEDLINE database (January 1950 to April 2009).
Intervention: Intensive lowering of blood glucose regimens (varied). Trials were chosen in which patients were randomly assigned to a treatment regimen that aimed to produce a significantly lower level of HbA1c compared with the standard regimen. They included trials such as ADVANCE, ACCORD which added additional drugs to achieve a target HbA1c of <6% or <6.5%, UKPDS where a target blood glucose level was aimed for and VADT where a reduction in HbA1c of 1.5% was the aim.
Comparison: Standard regimen (placebo, standard care, or glycaemic control of reduced intensity).
Outcomes: CV disease, coronary heart disease, stroke (fatal and nonfatal), CHF, MI (fatal and nonfatal), CV disease mortality, all-cause mortality, and severe hypoglycaemia.
Results: Results are presented for the random-effects models because there were differences between studies for several outcomes, as well as for duration of diabetes, achieved HbA1c and therapeutic regimens. Intensive glucose control reduced the risk of CV disease (RR 0.90, 95%CI 0.83 to 0.98; risk difference [per 1000 patients over 5 years of treatments] –15, 95%CI –24 to –5) but not CV death (RR 0.97, 95%CI 0.76 to 1.24; risk difference –3, 95%CI –14 to 7) or all-cause mortality (RR 0.98, 95%CI 0.84 to 1.15; risk difference –4, 95%CI –17 to 10). However, the risk of severe hypoglycaemia was doubled with intensive treatment (RR 2.03 95%CI 1.46 to 2.81; risk difference 39, 95%CI 7 to 71). There were also statistically significant reductions in coronary heart disease (RR 0.89, 95%CI 0.81 to 0.96; risk difference –11, 95%CI –17 to –5) and non-fatal MI (RR 0.84, 95%CI 0.75 to 0.94; risk difference –9, 95%CI –16 to 3) for patients on intensive therapy compared with standard treatment, but fatal MI (RR 0.94, 95%CI 0.75 to 1.18; risk difference –3, 95%CI –10 to 4), nonfatal stroke (RR 0.98, 95%CI 0.82 to 1.17; risk difference –3, 95%CI –7 to 2), fatal stroke (RR 0.87, 95%CI 0.63 to 1.20; risk difference 0, 95%CI –2 to 1), CHF (RR 1.01, 95%CI 0.89 to 1.14; risk difference 0, 95%CI –5 to 4) and peripheral arterial disease (RR 0.91, 95%CI 0.79 to 1.03; risk difference –3, 95%CI –5 to –1) did not differ between groups. These results are from pooling the summary findings from trials, rather than using individual data from patients. Possible heterogeneity was identified between the results of some subgroup analyses between the different trials for the effect of intensive glucose control on CV mortality and all-cause mortality.
The duration of trials ranged from 3.4 to 10.7 years. All trials showed lower postintervention HbA1c levels in the intensive than in the conventional glucose control group, with median differences ranging from –0.5% to –1.4%. The sample size-weighted overall difference in median HbA1c was –0.8%.
Sponsorship: This study was funded in part by career development awards from the National Heart Lung, and Blood Institute and from the Eunice Kennedy Shriver National Institute of Child Health and Human Development in the US.
Make sure you are signed up to NPC Email updates — the free email alerting system that keeps you up to date with the NPC news and outputs relevant to you.