22 January 2010
A Canadian case-control study has found an association between low HbA1c and risk of car crashes in people with diabetes. They found a relative increase in risk of a crash of about 26% for every 1% (11mmol/mol) lower HbA1c. Risk of a crash was about four times greater in people with a history of severe hypoglycaemia requiring outside assistance.
Level of evidence:
Level 2 (limited quality patient-oriented evidence) according to the SORT criteria.
Health professionals should follow NICE guidance on the management of type 2 diabetes. They state that individual targets for HbA1c should be agreed with each patient and that these could be greater than the general target of 6.5%(48mmol/mol). Highly intensive management to levels of less than 6.5% (48mmol/mol) should be avoided. While this study has several limitations, and association does not necessarily mean causation, they provide another consideration in agreeing HbA1c targets with individuals alongside patient preferences, the balance of likely benefits and harms, and the burden of extra medication.
What is the background to this?
As we have blogged previously, it remains uncertain whether intensive glucose control (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 (including smoking cessation, exercise, losing weight, controlling blood pressure, taking statins, etc).
In the UK, the rules for motor vehicle drivers require people with diabetes to inform the DVLA if they are treated with insulin, or have a history of certain problems relating to hypoglycaemia (and also in certain other circumstances). Similar rules apply in other countries. In this case control study the authors compared people who had had a car crash (cases) with those who had not (controls) among 795 people with diabetes and a record of HbA1c reporting to the Ontario driver licensing authority.
What does this study claim?
The mean HbA1c was lower among those who had had crashes than controls (7.4% vs 7.9%, 57mmol/mol vs 63mmol/mol, P=0.019). The odds ratio (OR) associated with a 1% (11mmol/mol) lower HbA1c was 1.26 (95% confidence interval [CI]1.03 to 1.54).
The risk among those in the lowest HbA1c quartile (4.4% to 6.9% [25mmol/mol to 52mmol/mol]) was more than double that of those in the highest quartile (9.0% to 14.7% [75mmol/mol to 137mmol/mol]). In absolute terms, this amounted to an extra 29 crashes (95%CI 16 to 46) compared to the number that would have occurred had all people had the same risk as those in the highest HbA1c quartile: about half the total number of crashes. The observed difference persisted when adjusted for potential confounders.
In addition, a history of severe hypoglycaemia which required outside assistance was associated with a substantially higher risk of a crash (OR 4.07, 95%CI 2.35 to 7.04). Older age of diabetes diagnosis was also associated with a greater risk (expressed as risk per decade, OR 1.29, 95%CI 1.07 to 1.57).
A previous blog points out the conflicting nature of evidence from randomised controlled trials (RCTs), with regard to the benefits and risks of intensive glucose control. The ACCORD, ADVANCE and VADT RCTs have not consistently identified a significant benefit from intensive glycaemic control in the treatment of type 2 diabetes with regard to CV outcomes and mortality (see also blogs 258, 147 and 64 for more details). Indeed, the ACCORD study was stopped early because of an increased risk of death in the intensive treatment arm.
This observational study has several limitations. The controls were not a random sample of the population because they came to attention through reports submitted by others or because of legal requirements for having a valid driver’s license. Controls ought to include all diabetic drivers who developed diabetes or obtained a license during the study period, but do not because of non-compliance with legislation or other reasons. We do not know how the rate of car crashes among people in Ontario with diabetes compares to those who do not have diabetes, or other relevant medical conditions, so as to put the difference in rates observed here into wider perspective.
Moreover, in ‘real life’, treatments are chosen, changed, or actively not chosen in the light of individual patients’ circumstances. Thus observed differences in outcomes may well be due to differences among the patients, not only the different treatments. Observational studies attempt to adjust for these differences by statistical modelling, but this requires certain assumptions and of course can take into account only those factors that have been recognised. Ideally, hypotheses generated from observational studies should be tested in RCTs. However, it would seem unlikely to be either ethical or practicable to do so in this case.
Despite the limitations, the results of this study add further grounds to review carefully the pursuit of very low HbA1c and supports NICE’s recommendation to
agree individual HbA1c targets with patients.
Design Case control study
Patients All drivers reported to the Ontario Ministry of Transportation Medical Advisory Board who had an underlying diagnosis of diabetes mellitus (795 out of a total of 3,900). This included all licensed drivers in Ontario with the accrual interval spanning from January 1, 2005 to January 1, 2007. Candidates were identified from mandatory annual reviews submitted by drivers who held commercial licenses or mandatory reports submitted in the aftermath of a documented motor vehicle crash, and all other diabetic patients reviewed for any other reason. Individuals were excluded if no HbA1c was available.
Cases and controls Individuals involved in a motor vehicle crash were defined as cases. Such cases were identified by the authorities responsible for investigating a crash. All other individuals who were not involved in a motor vehicle crash were defined as controls. Note: such controls are not a random sample of the population because they come to attention by reports submitted by others or because of legal requirements for having a valid driver’s license, see above. The mean age of cases was 50 years, 77% were male and 82% were receiving insulin. The mean age of controls was 52 years, 85% were male and 80% were receiving insulin.
Outcomes and results Mean HbA1c was lower among cases than controls (7.4% vs 7.9%, 57mmol/mol vs 63mmol/mol). The OR associated with a 1% (11mmol/mol) lower HbA1c was 1.26 (95%CI 1.03 to 1.54). The risk among those in the lowest HbA1c quartile (4.4% to 6.9% [25mmol/mol to 52
mmol/mol]) was more than double that of those in the highest quartile (9.0% to 14.7% [75mmol/mol to 137mmol/mol]). In absolute terms, this amounted to an extra 29 crashes (95%CI 16 to 46) compared to the number that would have occurred had all people had the same risk as those in the highest HbA1c quartile: about half the total number of crashes. The observed difference persisted when adjusted for potential confounders.
A history of severe hypoglycaemia which required outside assistance was associated with a substantially higher risk of a crash (OR 4.07, 95%CI 2.35 to 7.04). Older age of diabetes diagnosis was also associated with a greater risk (expressed as risk per decade, OR 1.29, 95%CI 1.07 to 1.57).
Sponsorship This project was supported by the Canada Research Chair in Medical Decision Sciences. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
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.