The recent subgroup analysis by Sämann et al. (1) examined the effectiveness of diabetes treatment and teaching programs (DTTPs) and advocated intensive insulin therapy combined with increased dietary freedom in individuals with type 1 diabetes who were at increased risk of hypoglycemia. There were significant reductions in hypoglycemic episodes, from 6.1 to 1.4 hypoglycemic events per patient per year, after participation. Sämann et al. concluded that DTTPs may reduce the occurrence of hypoglycemic episodes for those at risk, but, for several reasons, this conclusion is questionable.
As the authors noted, regression to the mean explained at least part of the observed reductions, since a subgroup with higher previous incidence of hypoglycemic episodes was analyzed. From statistical simulations of baseline results, a reduction from 6.1 to 4.7 (95% CI ±0.2) hypoglycemic events per patient per year may be expected simply due to regression to the mean. Furthermore, there was selection during recruitment of the whole study cohort, as patients frequently experiencing hypoglycemia were preferentially referred from their general practitioners or diabetologists (2). Regression to the mean would therefore explain an even greater reduction in the subgroup from 6.1 to <4.7 hypoglycemic events per patient per year.
As Sämann et al. also noted, interviewer bias may have been a particular problem. Baseline recall was purely based on participant’s memory of the previous year, which would be vulnerable to interviewer influence. Patients were better prepared for their 1-year follow-up but must have felt some pressure to report improvements to the research team, who were to present results at annual meetings.
It is difficult to distinguish to what extent hypoglycemic episodes were decreased by simply increasing the frequency of blood glucose testing. As stated by Sämann et al. (1), this clearly improves glucose control (3), and there may have been little additional improvement on hypoglycemia by DPPTs.
A group at high risk of hypoglycemia, the young and physically active (4), are not well represented in this cohort, and, as with other trials of this type (5,6), the average age of participants (38 years) was relatively high. In a younger age-group (mean age 27 years), the occurrence of hypoglycemic episodes increased threefold, while following an intensive insulin program (7) and a less intensive approach during exercise has recently been recommended (8) for younger individuals with type 1 diabetes.
In light of the currently available information, controlled studies of the impact of DTTPs on hypoglycemia are needed; this is especially so for the young and physically active.
