Abstract
S
In the current issue, Boulet et al. 6 report on the prevalence of insulin pump use in a very interesting group of 305 Canadian patients whose duration of type 1 diabetes is >50 years. Data were collected using a detailed mailed questionnaire. Forty-four percent of the respondents reported using a pump for a median of 8 years. Twenty of the 133 pump users had been using a pump for more than 20 years. Pump users were slightly younger and better educated than the nonpump group. Although only a few of the subjects habitually used continuous glucose monitoring (CGM), rates were higher among the pump group, 11% versus 5%.
Although minor hypoglycemia events (self-monitoring of blood glucose (SMBG) <4 mmol/L) occurred significantly more often among the pump users (5.1 vs. 6.5 events per month), severe hypoglycemia was significantly reduced among the pump users (1.3 vs. 0.5 events in the past year). Habitual CGM use was likewise associated with a substantially lower rate of severe hypoglycemia. The mean HbA1cs were not statistically different; 7.4% in the pump group versus 7.6% in the nonpump group. In multivariate analysis, only the daily frequency glucose tests and rate of minor hypoglycemia were associated with a lower HbA1c. Intriguingly, each incremental daily SMBG was associated with a 0.12% decrease in HbA1c.
On the other end of the age spectrum, Campbell et al. 7 used T1D Exchange data to compare 2684 children and adolescents in poor control (HbA1c > 8.9%) with 588 in excellent control (HbA1c < 7%). They found a multitude of patient characteristics and diabetes management practices were associated with membership in the excellent control group. As an example, adjusting for other factors, those who never missed an insulin bolus, compared with those who missed three or more boluses weekly, were 24.5 times more likely to be in the excellent control group. Similarly adjusted, a parallel analysis showed that those using pumps were 3.1 times more likely to be in excellent control.
Do Pumps Improve Diabetes Control?
Systematic reviews and meta-analysis of the impact of pump therapy on diabetes management have generally shown similar results, a somewhat lower HbA1c and often a reduced rate of hypoglycemia among those on pumps.
Reviewing 41 studies comprehensively, Golden et al.
8
concluded that: “both CSII and MDI have similar effectiveness on glycemic control and severe hypoglycemia, except in adults with type 1 diabetes where CSII had a favorable effect on HbA1c.” “real-CGM is superior to SMBG in lowering HbA1c, without affecting the risk of severe hypoglycemia, …… particularly when compliance is high.” “sensor-augmented pumps are superior to MDI/SMBG in lowering HbA1c.”
A recent review from the Canadian Agency for Drugs and Technologies in Health 9 concluded, in part, that “clinical effectiveness of CSII versus MDI in adult patients or in pregnant women with type 1 diabetes remains uncertain. Insulin pumps with integrated CGM (SAPT) appear to have better glycemic control without increasing the risk of hypoglycemia compared with MDI.”
Do Diabetes Technology Systems Improve Diabetes Control?
There are many challenges associated with the evaluation of diabetes technology and thus with the meta-analysis of past randomized controlled trials such as the previous examples. First, diabetes-related technologies are evolving rapidly and meta-analyses perforce must look at studies completed and published, often years ago. The quality and reliability of glucose sensors are objectively improving; anecdotally the clinical persistence of sensor use is improving. Early versions of the artificial pancreas, systems with threshold, or predictive low-glucose suspension of insulin infusion are now available and are effective in reducing nocturnal hypoglycemia in adults 10 and children. 11 Multiple studies of experimental full closed loop systems have demonstrated both safety and efficacy. 12,13 The first commercially available full closed loop artificial pancreas systems are expected quite soon.
Second, interventions involving glucose sensors and insulin infusion pump are challenging to test clinically. Potential subjects who really want the technology know that it is available and can be reluctant to enter studies. Buckingham et al., 14 as an example, reported that in a study of the effectiveness of early intensive therapy on beta-cell preservation in new onset type 1 diabetes, 15 out of 20 subjects in the usual care control group were using pumps within the first year.
Studies evolving technological systems such as the artificial pancreas present additional methodology challenges. Subjects who agree to enroll in early proof of principle studies tend to be technologically adept, potentially limiting the applicability of the results to the broader population of those with diabetes. They are frequently in excellent glycemic control, making it difficult to show an improvement in glycemic control. It is difficult to select the most appropriate control group. As an example, should a full closed loop artificial pancreas system be compared with multiple daily injections (MDI), CSII, or with a sensor-augmented predictive low-glucose suspend pump-sensor system? As such, I believe that systematic reviews of past studies underestimate the positive impact of diabetes-related technologies, including CSII, on patients with diabetes.
So, is pump therapy for all with type 1 diabetes? At this point, many benefit from pump therapy but some patients do well with MDI. With coming refinements and increasing availability of integrated artificial pancreas systems, essentially all those with type 1 diabetes will benefit from “enhanced” pump therapy.
Footnotes
Author Disclosure Statement
No competing financial interests exist.