Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

SMBG and glycaemic control

Study
Results
New users of SMBG
Prevalent users of SMBG
Comment

Bandolier is always on the lookout for useful papers on self monitoring of blood glucose, especially in those with type-2 diabetes. The ongoing controversy makes it interesting. A small number of small, inadequate, randomised trials say that self monitoring is not useful, yet this, and a few small observational studies hold sway over large, good, observational studies that point to substantial benefits. One such study in Germany featured in Bandolier 148.

We now have another large and impressive observational study with some unique features [1]. The results support common sense and experience over inadequate assessment of evidence – another fine example about why we need good evidence assessment.

Study


The study was designed to examine the association between changes in SMBG frequency and glycaemic control over four years. Two large cohorts were used to do this, identified from a diabetes registry with about 180,000 patients, part of a larger group of three million patients. Eligibility included members with continuous membership and full pharmacy benefits to avoid obvious problems if neither condition were met.

The first cohort had 16,000 patients who were not practising self monitoring of blood glucose at baseline. The longitudinal changes in HbA 1c were evaluated over four years, according to those who became users, and those who were persistent non-users.

The second cohort was 15,000 prevalent users who were practising self monitoring at baseline. This assessed how individual changes in frequency of self monitoring were reflected in HbA 1c over three years.

Patients were classified as being on no medication, oral hypoglycaemic agents, or insulin. Patients switching between therapies were excluded, so groups refer to those who were on that treatment only for the period of observation, though dose changes within a type of therapy could have occurred.

Results


The average age of patients in various groups was 53 to 67 years, with insulin users tending to be younger, and less likely to be female. Average initial HbA 1c concentrations also differed between groups. Changes in HbA 1c concentrations were adjusted for age, sex, comorbidity, baseline HbA 1c concentration, and a number of other relevant factors.

New users of SMBG


For those on no medication, new users had higher initial HbA 1c concentrations (8.2%) than persistent nonusers (6.6%). In the 7,800 new users, HbA 1c concentrations fell by larger amounts according to the number of strips per day used (Figure 1).


Figure 1: Association between change in strip use and change in HbA1c concentration over four years in patients on no medication






For those on oral hypoglycaemics only, new users had higher initial HbA 1c concentrations (8.6%) than persistent nonusers (7.3%). In the 5,500 new users, HbA 1c concentrations fell by larger amounts according to the number of strips per day used (Figure 2), up to three strips per day.


Figure 2: Association between change in strip use and change in HbA1c concentration over four years in patients on oral hypoglycaemics






Insulin users had high HbA 1c concentrations (9.3%), and in 840 of them there was a similar greater decrease in HbA 1c concentration with increasing strip use, by a maximum of about 0.8% with three strips per day or more.

Prevalent users of SMBG


In the prevalent cohort, those on no medication had the lowest HbA 1c concentration (6.4%), those on oral hypoglycaemic drugs only a higher HbA 1c concentration (7.6%), and those on insulin only the highest (8.1%). The average absolute change in frequency of strip use was 0.4, 0.5, and 1.1 strips per day respectively.

Among 1,600 patients on no medication, changes in strip frequency were not associated with change in HbA 1c concentrations. In those on oral hypoglycaemic drugs only (7,400) or insulin only (6,300), subsequent increases in strip use by one strip daily resulted in 0.16% and 0.12% reduction in HbA 1c concentrations respectively. There was a dose response for increasing use of strips.

Comment


These results are absolutely in line with what we have come to expect from large, good, observational studies related to self monitoring of blood glucose and glycaemic control as measured by HbA 1c concentrations. It says that increased use results in better glycaemic control, and (if you look at the data), that you get better results with those with a higher initial HbA 1c concentration. Common, sense, really.

Bandolier likes the advice on SMBG in type 2 diabetes given by an international committee in 2005 [2]. It said that individuals offered self monitoring of blood glucose on an ongoing basis might include recently diagnosed diabetics, those with more erratic lifestyles, people with problems with hypoglycaemia, and those keen to tighten their glucose control, which is what doctors and other professionals are very good at doing.

That seems more sensible than a blanket view of cost effectiveness based on flawed evidence, which is what we seem have in the UK. To recap, much weight has previously been given to a meta-analysis of tiny, invalid, randomised trials, and some small observational studies. Less weight seems to have been given to large, well-conducted, observational studies. It is a bit more complicated, but even a passing knowledge of diabetes combined with an appreciation of what constitutes good evidence leads to the obvious conclusion that SMBG is worthwhile in some patients.

References:

  1. AJ Karter et al. Longitudinal study of new and prevalent use of self-monitoring of blood glucose. Diabetes Care 2006 29: 1757-1763.
  2. Global guidelines for type 2 diabetes. International Diabetes Federation, 2005 (www.idf.org).

previous or next story