Insulin Pump After Failure of Multiple Daily Injections
Insulin Pump After Failure of Multiple Daily Injections
Despite recent attention, the role of CSII in DM2 is unclear. A recent retrospective claims analysis of over 900 patients found improved HbA1c in patients with DM2 switching from insulin to CSII. However, randomized controlled trials comparing CSII to multiple daily injections (MDI) have not shown a clear benefit of CSII. One small randomized controlled trial compared CSII to conventional insulin therapy and found CSII superior. Additional studies have been reported that do show benefits of CSII therapy in various settings in terms of improved glucose control and patient satisfaction. Two recent reports have evaluated moving to CSII in a small number of DM2 patients not at goal on MDI therapy. Both studies reported a significant lowering of A1c and a decrease in total insulin use.
We found evidence of improvement in glucose control in patients in our specialty diabetes clinic who were moved to CSII after failure of MDI, as judged by elevated HbA1c and/or hypoglycemia. The degree of improvement in HbA1c was greater for those patients who started with higher levels of glycemia (HbA1c 8.4% or greater). The improvement in glycemia was not due to increased amounts of administered insulin. In fact, while weight did increase (reflecting improved control), daily basal insulin requirements actually decreased when HgA1c was lowered. This could result from 2 distinct possibilities. One is an improvement in insulin sensitivity. This would not be expected without weight loss (much less with weight gain). When improvement in profound hyperglycemia is accompanied by enhanced insulin secretion and insulin responsiveness, the concept of "glucose toxicity" has been invoked. However, our patients did not have such profound hyperglycemia. Another possibility is simply that insulin was administered in a more efficient fashion. Hence, there was more appropriate matching of insulin administration with insulin need. This would, of course, be accompanied by fewer periods of hypoglycemia. We cannot address this question directly, since we did not assess rates of hypoglycemia.
Seven patients in this study were using U-500 insulin via CSII. Subgroup analysis demonstrated that they also had a significant decrease in HbA1C of 1.1%, together with a drop in basal insulin requirements. This is similar, in terms of improvement in HbA1c, to what Lane et al recently reported in a prospective study of 21 patients failing MDI put on CSII with U-500 insulin. They did not, however, notice a drop in total insulin dose.
There are numerous barriers to CSII use in patients with DM2. One is the level of complexity required for bolus administration. Software programs on pumps are useful tools in this regard, but these calculation programs require input of patient collected data (current glucose level and accurate carbohydrate content of the meal). Many insulin regimens avoid this by using fixed doses at meals. One question we asked was whether a patient's inability or unwillingness to use bolus calculation routinely (over 50% of the time) in the setting of CSII negated its benefits. We found this not to be the case: manual users had similar improvements in overall glucose control. Two other investigators have also reported that simplified CSII regimens can be effective. This implies that successful CSII candidates do not necessarily need to be well versed in carbohydrate counting or use frequent dose adjustment. This lack of overall superiority of calculated bolus administration could be due to residual insulin secretion in some patients.
There are some limitations to our study. First, this was not a randomized, controlled trial. There were no specific selection criteria for CSII, and thus a clearly defined picture was not available for who might benefit most. Conversely, the use of strict selection criteria lessens the applicability of the data to the general population. Hence, the "real world" nature of our population is both a limitation and strength of our study. We were unable to accurately determine total daily insulin doses from retrospective chart review. Our approach for patients with MDI or CSII is to have a roughly 50–50 split with basal and bolus insulin, so we believe that the basal doses do correlate well with total insulin. Another limitation of our study was that we did not assess hypoglycemia. This would have helped understand the benefits (or lack thereof) of CSII use in patients, especially those with baseline HbA1c close to goal. It would also have helped our understanding of the finding of decreased insulin dosing being associated with lower HbA1c.
Discussion
Despite recent attention, the role of CSII in DM2 is unclear. A recent retrospective claims analysis of over 900 patients found improved HbA1c in patients with DM2 switching from insulin to CSII. However, randomized controlled trials comparing CSII to multiple daily injections (MDI) have not shown a clear benefit of CSII. One small randomized controlled trial compared CSII to conventional insulin therapy and found CSII superior. Additional studies have been reported that do show benefits of CSII therapy in various settings in terms of improved glucose control and patient satisfaction. Two recent reports have evaluated moving to CSII in a small number of DM2 patients not at goal on MDI therapy. Both studies reported a significant lowering of A1c and a decrease in total insulin use.
We found evidence of improvement in glucose control in patients in our specialty diabetes clinic who were moved to CSII after failure of MDI, as judged by elevated HbA1c and/or hypoglycemia. The degree of improvement in HbA1c was greater for those patients who started with higher levels of glycemia (HbA1c 8.4% or greater). The improvement in glycemia was not due to increased amounts of administered insulin. In fact, while weight did increase (reflecting improved control), daily basal insulin requirements actually decreased when HgA1c was lowered. This could result from 2 distinct possibilities. One is an improvement in insulin sensitivity. This would not be expected without weight loss (much less with weight gain). When improvement in profound hyperglycemia is accompanied by enhanced insulin secretion and insulin responsiveness, the concept of "glucose toxicity" has been invoked. However, our patients did not have such profound hyperglycemia. Another possibility is simply that insulin was administered in a more efficient fashion. Hence, there was more appropriate matching of insulin administration with insulin need. This would, of course, be accompanied by fewer periods of hypoglycemia. We cannot address this question directly, since we did not assess rates of hypoglycemia.
Seven patients in this study were using U-500 insulin via CSII. Subgroup analysis demonstrated that they also had a significant decrease in HbA1C of 1.1%, together with a drop in basal insulin requirements. This is similar, in terms of improvement in HbA1c, to what Lane et al recently reported in a prospective study of 21 patients failing MDI put on CSII with U-500 insulin. They did not, however, notice a drop in total insulin dose.
There are numerous barriers to CSII use in patients with DM2. One is the level of complexity required for bolus administration. Software programs on pumps are useful tools in this regard, but these calculation programs require input of patient collected data (current glucose level and accurate carbohydrate content of the meal). Many insulin regimens avoid this by using fixed doses at meals. One question we asked was whether a patient's inability or unwillingness to use bolus calculation routinely (over 50% of the time) in the setting of CSII negated its benefits. We found this not to be the case: manual users had similar improvements in overall glucose control. Two other investigators have also reported that simplified CSII regimens can be effective. This implies that successful CSII candidates do not necessarily need to be well versed in carbohydrate counting or use frequent dose adjustment. This lack of overall superiority of calculated bolus administration could be due to residual insulin secretion in some patients.
There are some limitations to our study. First, this was not a randomized, controlled trial. There were no specific selection criteria for CSII, and thus a clearly defined picture was not available for who might benefit most. Conversely, the use of strict selection criteria lessens the applicability of the data to the general population. Hence, the "real world" nature of our population is both a limitation and strength of our study. We were unable to accurately determine total daily insulin doses from retrospective chart review. Our approach for patients with MDI or CSII is to have a roughly 50–50 split with basal and bolus insulin, so we believe that the basal doses do correlate well with total insulin. Another limitation of our study was that we did not assess hypoglycemia. This would have helped understand the benefits (or lack thereof) of CSII use in patients, especially those with baseline HbA1c close to goal. It would also have helped our understanding of the finding of decreased insulin dosing being associated with lower HbA1c.
Source...